CN101459755B - Visual processing device - Google Patents
Visual processing device Download PDFInfo
- Publication number
- CN101459755B CN101459755B CN2008101843841A CN200810184384A CN101459755B CN 101459755 B CN101459755 B CN 101459755B CN 2008101843841 A CN2008101843841 A CN 2008101843841A CN 200810184384 A CN200810184384 A CN 200810184384A CN 101459755 B CN101459755 B CN 101459755B
- Authority
- CN
- China
- Prior art keywords
- image
- region
- gray scale
- value
- mentioned
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
It is an object of the present invention to provide a visual processing device that achieves gradation processing that further enhances the visual effect. The visual processing device 1 is a visual processing device that performs gradation processing for each image region of an input signal IS, and is provided with an image partitioning portion 2, a gradation transformation curve derivation portion 10, and a gradation processing portion 5. The image partitioning portion 2 and the gradation transformation curve derivation portion 10 use a brightness histogram Hm of a wide area image region Em to create a gradation transformation curve Cm of the image region Pm. The gradation processing portion 5 performs gradation processing of that image region Pm based on the gradation transformation curve Cm that has been derived.
Description
Technical field
The present invention relates to a kind of visual processing apparatus, the visual processing apparatus that particularly a kind of gray scale of carrying out picture signal is handled.In addition, the invention still further relates to a kind of visual processing method, visual processing program and semiconductor device.
Background technology
As the visual processes of the picture signal of original image, known have living space to handle with gray scale handle.
Spatial manipulation is meant, uses the neighboring pixel as the object pixel of process object, the processing of objective for implementation pixel.In addition, use the picture signal be implemented spatial manipulation, the contrast of carrying out original image is strengthened, the technology of dynamic range (DR) compression etc. also is known.During contrast is strengthened, poor (the clear composition of image) of original image and blurred signal added to original image, carry out the sharpening of image.In the DR compression, from original image, deduct the part of blurred signal, carry out the compression of dynamic range.
Gray scale is handled and to be meant, and is irrelevant with the neighboring pixel of object pixel, uses table look-up (LUT) etc. to carry out the processing of the conversion of pixel value to each object pixel, is also referred to as gamma-corrected.For example, under the situation of strengthening contrast, use the LUT that the gray scale of the higher gray scale of occurrence frequency in the original image is strengthened, carry out the conversion of pixel value.Use the gray scale of LUT to handle, known have pair original image all to determine to use the gray scale of 1 LUT to handle (histogram equalization method), and to cutting apart each of a plurality of image-regions that original image obtains, determine and use gray scale processing (local histogram's equalization method) (for example opening 2000-57335 communique (the 3rd page, the 13rd figure～the 16th figure)) of LUT respectively with reference to the spy.
Use Figure 33～Figure 36, illustrate, determine and use the gray scale processing of LUT cutting apart each of a plurality of image-regions that original image obtains.
What represent among Figure 33 is, visual processing apparatus 300 of LUT is determined and used to each that cut apart a plurality of image-regions that original image obtains.Visual processing apparatus 300 has: will be divided into as the original image of input signal IS input a plurality of image-region Sm (1≤m≤n:n is the number of cutting apart of original image) image segmentation portion 301, each image-region Sm is derived the gray-scale transformation curve leading-out portion 310 of gray-scale transformation curve Cm and loads gray-scale transformation curve Cm and the gray scale handling part 304 of the output signal OS that gray scale handles has been carried out in output to each image-region Sm.Gray-scale transformation curve leading-out portion 310 is by the histogram generating unit 302 that generates lightness (lightness) the histogram Hm in each image-region Sm, constitute with the grey scale curve generating unit 303 that generates each image-region Sm corresponding gray scale conversion curve Cm according to the lightness histogram Hm that is generated.
Use Figure 34～Figure 36, the action of each one is described.Image segmentation portion 301 will be divided into a plurality of (n) image-region (with reference to Figure 34 (a)) as the original image that input signal IS is imported.Histogram generating unit 302 generates the lightness histogram Hm (with reference to Figure 35) of each image-region Sm.The distribution of the brightness value of the both full-pixel in each lightness histogram Hm, presentation video region S m.That is, in the lightness histogram Hm shown in Figure 35 (a)～(d), transverse axis is represented the lightness grade of input signal IS, longitudinal axis remarked pixel number.Grey scale curve generating unit 303 is according to the order of lightness accumulation lightness histogram Hm " pixel count ", with this accumulation curve as gray-scale transformation curve Cm (with reference to Figure 36).Among the gray-scale transformation curve Cm shown in Figure 36, transverse axis is represented the brightness value of the pixel of the image-region Sm among the input signal IS, and the longitudinal axis is represented the brightness value of the pixel of the image-region Sm among the output signal OS.Gray scale handling part 304 loads gray-scale transformation curve Cm, and according to gray-scale transformation curve Cm, the brightness value of the pixel of the image-region Sm among the conversion input signal IS.By like this, in each block, improve the slope of the higher gray scale of occurrence frequency, by like this, improve the contrast sense of each block.
In the histogram generating unit 302, according to the lightness histogram Hm generation gray-scale transformation curve Cm of the pixel in the image-region Sm.In order to generate the gray-scale transformation curve Cm that uses among the image-region Sm more rightly, need all not have to highlights (Gao Guang) from the dark portion (shade) of image with omitting, and need with reference to considerable pixel.Therefore, each image-region Sm can not be too little, also be cutting apart of original image several n can not be too big.Though cut apart several n because of picture material is different, on experience, use 4～16 the number of cutting apart.
Because each image-region Sm can not be too little, therefore among the output signal OS after gray scale is handled, produce following problem sometimes.Also promptly, handle owing to use 1 gray-scale transformation curve Cm to carry out gray scale to each image-region Sm, therefore the seam on the border of each image-region Sm is eye-catching completely naturally inadequately sometimes, produces false contouring in image-region Sm.In addition, cut apart number at most only 4～16, but image-region Sm is bigger, therefore exists under the situation of extremely different images between image-region, the deep or light variation between image-region is bigger, is difficult to prevent the generation of false contouring.For example shown in Figure 34 (b), Figure 34 (c),, deep or lightly change terrifically because of the position relation between image (for example the object in the image etc.) and the image-region Sm.
Summary of the invention
Therefore the visual processing apparatus that provides a kind of gray scale that can realize further improving visual effect to handle, is provided.
Visual processing apparatus described in the 1st invention of the present invention carries out gray scale according to each image-region to the picture signal of being imported and handles, and has greyscale transformation characteristic export agency and gray scale processing mechanism.Greyscale transformation characteristic export agency, it uses the peripheral view data of at least one ambient image regions to derive the greyscale transformation characteristic of the object image area of the object that becomes the gray scale processing, wherein, above-mentioned ambient image regions is the image-region that is positioned at the periphery of above-mentioned object image area, comprises a plurality of pixels.The gray scale processing mechanism, it is implemented gray scale to the picture signal of object image area and handles according to the greyscale transformation characteristic that is derived.
Object image area for example is by the pixel that is included in the picture signal, or picture signal is divided into zone that a plurality of pixel constituted outside the image block behind the given unit etc.Ambient image regions for example is picture signal to be divided into zone that a plurality of pixel constituted outside the image block behind the given unit etc.The periphery view data is the view data of ambient image regions, or from the data that view data derived, for example the pixel value of ambient image regions, gamma characteristic (brightness of each pixel or lightness), thumbnail (downscaled images or reduced resolution between get image) etc.In addition, ambient image regions can be positioned at the periphery of object image area, needs not be the zone that surrounds object image area.
In the visual processing apparatus of the present invention, when judging the greyscale transformation characteristic of object image area, use the peripheral view data of ambient image regions to judge.Therefore, handle the effect of adding spatial manipulation can for the gray scale of each object image area, thereby can realize further having improved the gray scale processing of visual effect.
Visual processing apparatus described in the 2nd invention of the present invention is a kind of according to the described visual processing apparatus of the 1st invention of the present invention, and ambient image regions is that picture signal is divided into the image block that obtains behind the given unit.
Here, image block is meant picture signal is divided into rectangle each zone afterwards.
In the visual processing apparatus of the present invention, can handle ambient image regions with image block unit.Therefore, can reduce the decision of ambient image regions and the needed processing load of derivation of greyscale transformation characteristic.
Visual processing apparatus described in the 3rd invention of the present invention, be a kind of according to the described visual processing apparatus of the of the present invention the 1st or 2 inventions, greyscale transformation characteristic export agency also uses the object image data of object image area, the greyscale transformation characteristic of coming the derived object image-region.
Object image data is the view data of object image area, or from the data that view data derived, for example the pixel value of object image area, gamma characteristic (brightness of each pixel or lightness), thumbnail (downscaled images or reduced resolution between get image) etc.
In the visual processing apparatus of the present invention, when judging the greyscale transformation characteristic of object image area,, also use the peripheral view data of ambient image regions to judge except the object image data of object image area.Therefore, handle the effect of adding spatial manipulation can for the gray scale of object image area, thereby can realize further having improved the gray scale processing of visual effect.
Visual processing apparatus described in the 4th invention of the present invention, be a kind of according to the described visual processing apparatus of the 3rd invention of the present invention, greyscale transformation characteristic export agency, have: the characteristic parameter export agency, use object image data and peripheral view data to derive characteristic parameter, this characteristic parameter is the parameter of the feature of indicated object image-region; And greyscale transformation characteristic determination means, the characteristic parameter of the subject area that is derived according to the characteristic parameter export agency, decision greyscale transformation characteristic.
Characteristic parameter for example is mean value (simple mean value, weighted average etc.), typical value (maximum, minimum value, median etc.) or the histogram etc. of object image data and peripheral view data etc.Here, histogram is meant the distribution of the gamma characteristic of object image data for example and peripheral view data.
In the visual processing apparatus of the present invention,, also use peripheral view data to derive characteristic parameter except object image data.Therefore, handle the effect of adding spatial manipulation can for the gray scale of object image area, thereby can realize further having improved the gray scale processing of visual effect.More specifically effect is, can suppress the generation that gray scale is handled caused false contouring.In addition, the border that can prevent object image area becomes obvious artificially.
Visual processing apparatus described in the 5th invention of the present invention is a kind of according to the described visual processing apparatus of the 4th invention of the present invention, it is characterized in that characteristic parameter is a histogram.
Greyscale transformation characteristic determination means, the cumulative curve decision that for example will accumulate histogrammic value is the greyscale transformation characteristic, or selects corresponding histogrammic greyscale transformation characteristic.
In the visual processing apparatus of the present invention,, also use peripheral view data to generate histogram except object image data.Therefore, can suppress the generation that gray scale is handled caused false contouring.In addition, the border that can prevent object image area becomes obvious artificially.
Of the present invention the 6th the invention described in visual processing apparatus, be a kind of according to of the present invention the 4th the invention described visual processing apparatus, it is characterized in that, greyscale transformation characteristic determination means, the use characteristic parameter, select form change in advance the greyscale transformation characteristic.
Here, the data that the greyscale transformation characteristic is the form change are preserved the characteristic of the object image data after the object image data corresponding gray scale is handled in the form.
Greyscale transformation characteristic determination means is selected the pairing form of value of each characteristic parameter.
In the visual processing apparatus of the present invention, use the form change the greyscale transformation characteristic carry out gray scale and handle.Therefore can allow gray scale handle high speed.In addition, carry out the gray scale processing, therefore can carry out suitable gray scale and handle owing to from a plurality of forms, select 1 form.
Of the present invention the 7th the invention described in visual processing apparatus, be a kind of according to of the present invention the 6th the invention described visual processing apparatus, be characterised in that, form change in advance the greyscale transformation characteristic variable.
In the visual processing apparatus of the present invention,, do not need to change hardware configuration, just can carry out various changes the characteristic that gray scale is handled by change greyscale transformation characteristic.
Visual processing apparatus described in the 8th invention of the present invention is a kind of according to the described visual processing apparatus of the 7th invention of the present invention, is characterised in that the change of greyscale transformation characteristic is by realizing at least a portion correction of greyscale transformation characteristic.
In the visual processing apparatus of the present invention,, carry out the change of greyscale transformation characteristic by at least a portion of greyscale transformation characteristic is revised.Therefore, can cut down the memory capacity that is used for the greyscale transformation characteristic, realize various gray scales processing simultaneously.
Visual processing apparatus described in the 9th invention of the present invention is a kind ofly to be characterised in that according to the described visual processing apparatus of the 4th invention of the present invention, greyscale transformation characteristic determination means, and the use characteristic parameter generates the greyscale transformation characteristic by the computing that is predetermined.
Here, the greyscale transformation characteristic provides the object image data after the object image data corresponding gray scale is handled.In addition, generate the computing of greyscale transformation characteristic, the use characteristic parameter pre-determines.In more detail, for example, the corresponding computing of value of selection and each characteristic parameter, perhaps the value according to characteristic parameter generates computing.
In the visual processing apparatus of the present invention, do not need to preserve in advance the greyscale transformation characteristic, thereby can cut down the memory capacity that is used for preserving the greyscale transformation characteristic.
Visual processing apparatus described in the 10th invention of the present invention is a kind of according to the described visual processing apparatus of the 9th invention of the present invention, is characterised in that the computing that is predetermined is variable.
Visual processing apparatus of the present invention by the change computing, can carry out various changes to the characteristic that gray scale is handled.
Visual processing apparatus described in the 11st invention of the present invention is a kind of according to the described visual processing apparatus of the 10th invention of the present invention, is characterised in that the change of computing realizes by at least a portion correction to computing.
Visual processing apparatus of the present invention is revised by at least a portion to computing, changes the greyscale transformation characteristic.Therefore, even it is identical to be used for preserving the memory capacity of computing, can realize that also more kinds of gray scales handles.
Visual processing apparatus described in the 12nd invention of the present invention is a kind of according to the described visual processing apparatus of the 4th invention of the present invention, is characterised in that the greyscale transformation characteristic obtains by interpolation or a plurality of greyscale transformation characteristics of extrapolation.
Here, greyscale transformation characteristic for example is the characteristic of the object image data after the object image data corresponding gray scale is handled.The greyscale transformation characteristic provides by for example form exclusive disjunction form.
Visual processing apparatus of the present invention can use by interpolation or the resulting new greyscale transformation characteristic of a plurality of greyscale transformation characteristics of extrapolation, carries out gray scale and handles.Therefore, even cut down the memory capacity that is used to preserve the greyscale transformation characteristic, also can realize more kinds of gray scale processing.
Visual processing method described in the 13rd invention of the present invention carries out gray scale according to each image-region to the picture signal of being imported and handles, and having the greyscale transformation characteristic derives step, and the gray scale treatment step.The greyscale transformation characteristic derives step, it uses the peripheral view data of at least one ambient image regions to derive the greyscale transformation characteristic of the object image area of the object that becomes above-mentioned gray scale processing, wherein, above-mentioned ambient image regions is the image-region that is positioned at the periphery of above-mentioned object image area, comprises a plurality of pixels.The gray scale treatment step, it is implemented gray scale to the picture signal of object image area and handles according to the greyscale transformation characteristic that is derived.
In the visual processing method of the present invention, when judging the greyscale transformation characteristic of object image area, use the peripheral view data of ambient image regions to judge.Therefore, handle the effect of adding spatial manipulation can for the gray scale of each object image area, thereby can realize further having improved the gray scale processing of visual effect.
Visual processing method described in the 14th invention of the present invention is a kind of according to the described visual processing method of the 13rd invention of the present invention, and ambient image regions is that picture signal is divided into the image block that obtains behind the given unit.
Visual processing method of the present invention can be handled ambient image regions with image block unit.Therefore, can reduce the decision of ambient image regions and the needed processing load of derivation of greyscale transformation characteristic.
Visual processing method described in the 15th invention of the present invention, be a kind of according to the described visual processing method of the of the present invention the 13rd or 14 inventions, the greyscale transformation characteristic derives step, also uses the object image data of object image area, the greyscale transformation characteristic of coming the derived object image-region.
Visual processing method of the present invention when judging the greyscale transformation characteristic of object image area, except the object image data of object image area, also uses the peripheral view data of ambient image regions to judge.Therefore, handle the effect of adding spatial manipulation can for the gray scale of object image area, thereby can realize further having improved the gray scale processing of visual effect.
Visual processing method described in the 16th invention of the present invention, be a kind of according to the described visual processing method of the 15th invention of the present invention, the greyscale transformation characteristic derives step, have: characteristic parameter is derived step, use object image data and peripheral view data to derive characteristic parameter, this characteristic parameter is the parameter of the feature of indicated object image-region; And greyscale transformation characteristic deciding step, according to the characteristic parameter that characteristic parameter is derived the subject area that step derived, decision greyscale transformation characteristic.
Visual processing method of the present invention except object image data, also uses peripheral view data to derive characteristic parameter.Therefore, handle the effect of adding spatial manipulation can for the gray scale of object image area, thereby can realize further having improved the gray scale processing of visual effect.More specifically effect is, can suppress the generation that gray scale is handled caused false contouring.In addition, the border that can prevent object image area becomes obvious artificially.
Visual processing program described in the 17th invention of the present invention is used for using a computer and implements according to each image-region the picture signal of being imported to be carried out the visual processing method that gray scale is handled.This visual processing method, having the greyscale transformation characteristic derives step and gray scale treatment step.The greyscale transformation characteristic derives step, it uses the peripheral view data of at least one ambient image regions to derive the greyscale transformation characteristic of the object image area of the object that becomes above-mentioned gray scale processing, wherein, above-mentioned ambient image regions is the image-region that is positioned at the periphery of above-mentioned object image area, comprises a plurality of pixels.The gray scale treatment step, it is implemented gray scale to the picture signal of object image area and handles according to the greyscale transformation characteristic that is derived.
Visual processing program of the present invention when judging the greyscale transformation characteristic of object image area, uses the peripheral view data of ambient image regions to judge.Therefore, handle the effect of adding spatial manipulation can for the gray scale of each object image area, thereby can realize further having improved the gray scale processing of visual effect.
Visual processing program described in the 18th invention of the present invention is a kind of according to the described visual processing program of the 17th invention of the present invention, and ambient image regions is that picture signal is divided into the image block that obtains behind the given unit.
Visual processing program of the present invention can be handled ambient image regions with image block unit.Therefore, can reduce the decision of ambient image regions and the needed processing load of derivation of greyscale transformation characteristic.
Visual processing program described in the 19th invention of the present invention, be a kind of according to the described visual processing program of the of the present invention the 17th or 18 inventions, the greyscale transformation characteristic derives step, also uses the object image data of object image area, the greyscale transformation characteristic of coming the derived object image-region.
Visual processing program of the present invention when judging the greyscale transformation characteristic of object image area, except the object image data of object image area, also uses the peripheral view data of ambient image regions to judge.Therefore, handle the effect of adding spatial manipulation can for the gray scale of object image area, thereby can realize further having improved the gray scale processing of visual effect.
Visual processing program described in the 20th invention of the present invention, be a kind of according to the described visual processing program of the 19th invention of the present invention, the greyscale transformation characteristic derives step, have: characteristic parameter is derived step, use object image data and peripheral view data to derive characteristic parameter, this characteristic parameter is the parameter of the feature of indicated object image-region; And greyscale transformation characteristic deciding step, according to the characteristic parameter that characteristic parameter is derived the subject area that step derived, decision greyscale transformation characteristic.
Visual processing program of the present invention except object image data, also uses peripheral view data to derive characteristic parameter.Therefore, handle the effect of adding spatial manipulation can for the gray scale of object image area, thereby can realize further having improved the gray scale processing of visual effect.More specifically effect is, can suppress the generation that gray scale is handled caused false contouring.In addition, the border that can prevent object image area becomes obvious artificially.
Semiconductor device described in the 21st invention of the present invention carries out gray scale according to each image-region to the picture signal of being imported and handles, and has greyscale transformation characteristic leading-out portion and gray scale handling part.Greyscale transformation characteristic leading-out portion, it uses the peripheral view data of at least one ambient image regions to derive the greyscale transformation characteristic of the object image area of the object that becomes above-mentioned gray scale processing, wherein, above-mentioned ambient image regions is the image-region that is positioned at the periphery of above-mentioned object image area, comprises a plurality of pixels.The gray scale handling part, it is implemented gray scale to the picture signal of object image area and handles according to the greyscale transformation characteristic that is derived.
In the semiconductor device of the present invention, when judging the greyscale transformation characteristic of object image area, use the peripheral view data of ambient image regions to judge.Therefore, handle the effect of adding spatial manipulation can for the gray scale of each object image area, thereby can realize further having improved the gray scale processing of visual effect.
Semiconductor device described in the 22nd invention of the present invention is a kind of according to the described semiconductor device of the 21st invention of the present invention, and ambient image regions is that picture signal is divided into the image block that obtains behind the given unit.
Semiconductor device of the present invention can be handled ambient image regions with image block unit.Therefore, can reduce the decision of ambient image regions and the needed processing load of derivation of greyscale transformation characteristic.
Semiconductor device described in the 23rd invention of the present invention, be a kind of according to the described semiconductor device of the of the present invention the 21st or 22 inventions, greyscale transformation characteristic leading-out portion also uses the object image data of object image area, the greyscale transformation characteristic of coming the derived object image-region.
In the semiconductor device of the present invention, when judging the greyscale transformation characteristic of object image area,, also use the peripheral view data of ambient image regions to judge except the object image data of object image area.Therefore, handle the effect of adding spatial manipulation can for the gray scale of object image area, thereby can realize further having improved the gray scale processing of visual effect.
Semiconductor device described in the 24th invention of the present invention, be a kind of according to the described semiconductor device of the 23rd invention of the present invention, greyscale transformation characteristic leading-out portion, have: the characteristic parameter leading-out portion, use object image data and peripheral view data to derive characteristic parameter, this characteristic parameter is the parameter of the feature of indicated object image-region; And greyscale transformation characteristic determination section, the characteristic parameter of the subject area that is derived according to the characteristic parameter leading-out portion, decision greyscale transformation characteristic.
Semiconductor device of the present invention except object image data, also uses peripheral view data to derive characteristic parameter.Therefore, handle the effect of adding spatial manipulation can for the gray scale of object image area, thereby can realize further having improved the gray scale processing of visual effect.More specifically effect is, can suppress the generation that gray scale is handled caused false contouring.In addition, the border that can prevent object image area becomes obvious artificially.
By visual processing apparatus of the present invention, can realize further having improved the gray scale processing of visual effect.
Description of drawings
Fig. 1 is the block diagram (the 1st execution mode) of the structure of explanation visual processing apparatus 1.
Fig. 2 is the schematic diagram (the 1st execution mode) of explanation image-region Pm.
Fig. 3 is the schematic diagram (the 1st execution mode) of explanation lightness histogram Hm.
Fig. 4 is the schematic diagram (the 1st execution mode) of explanation gray-scale transformation curve Cm.
Fig. 5 is the flow chart (the 1st execution mode) of explanation visual processing method.
Fig. 6 is the block diagram (the 2nd execution mode) of the structure of explanation visual processing apparatus 11.
The schematic diagram (2nd execution mode) of Fig. 7 for gray-scale transformation curve candidate G1～Gp is described.
The schematic diagram (2nd execution mode) of Fig. 8 for two-dimentional LUT41 is described.
The schematic diagram (the 2nd execution mode) that Fig. 9 describes for the action to gray scale correction portion 15.
Figure 10 is the flow chart (the 2nd execution mode) of explanation visual processing method.
The schematic diagram (the 2nd execution mode) that Figure 11 describes for the variation to the selection of gray-scale transformation curve Cm.
The schematic diagram (the 2nd execution mode) that Figure 12 handles as the gray scale of variation for explanation.
Figure 13 is the block diagram (the 2nd execution mode) of the structure of explanation gray scale processing execution portion 44.
Figure 14 is the schematic diagram (the 2nd execution mode) of the relation between illustrative graph parameter P1 and P2 and the gray-scale transformation curve candidate G1～Gp.
Figure 15 is the schematic diagram (the 2nd execution mode) of the relation between illustrative graph parameter P1 and P2 and the selection signal Sm.
Figure 16 is the schematic diagram (the 2nd execution mode) of the relation between illustrative graph parameter P1 and P2 and the selection signal Sm.
Figure 17 is the schematic diagram (the 2nd execution mode) of the relation between illustrative graph parameter P1 and P2 and the gray-scale transformation curve candidate G1～Gp.
Figure 18 is the schematic diagram (the 2nd execution mode) of the relation between illustrative graph parameter P1 and P2 and the selection signal Sm.
Figure 19 is the block diagram (the 3rd execution mode) of the structure of explanation visual processing apparatus 21.
Figure 20 selects the schematic diagram (the 3rd execution mode) of the action of signal correction portion 24 for explanation.
Figure 21 is the flow chart (the 3rd execution mode) of explanation visual processing method.
Figure 22 is the block diagram (the 4th execution mode) of the structure of explanation visual processing apparatus 61.
The schematic diagram (the 4th execution mode) that Figure 23 describes for the spatial manipulation to spatial manipulation portion 62.
Figure 24 is the table (the 4th execution mode) of explanation weight coefficient [Wij].
Figure 25 is the schematic diagram (the 4th execution mode) of the effect of the visual processes of explanation visual processing apparatus 61.
Figure 26 is the block diagram (the 4th execution mode) of the structure of explanation visual processing apparatus 961.
Figure 27 is the schematic diagram (the 4th execution mode) of the spatial manipulation of explanation spatial manipulation portion 962.
Figure 28 is the table (the 4th execution mode) of explanation weight coefficient [Wij].
Figure 29 is the block diagram (the 6th execution mode) of all formations of description feed system.
Figure 30 is the example (the 6th execution mode) of the mobile phone of installation visual processing apparatus of the present invention.
Figure 31 is the block diagram (the 6th execution mode) of the formation of explanation mobile phone.
Figure 32 is the example (6th execution mode) of digital broadcasting with system.
Figure 33 is the block diagram (background technology) of the structure of explanation visual processing apparatus 300.
Figure 34 is the schematic diagram (background technology) of key diagram as region S m.
Figure 35 is the schematic diagram (background technology) of explanation lightness histogram Hm.
Figure 36 is the schematic diagram (background technology) of explanation gray-scale transformation curve Cm.
Embodiment
[the 1st execution mode]
Contrast Fig. 1～Fig. 5 describes the visual processing apparatus 1 as the 1st execution mode of the present invention.Visual processing apparatus 1 for example is built-in or be connected computer, television set, digital camera, mobile phone, PDA etc. handles in the machine of image, carries out the device of the gray scale processing of image.Visual processing apparatus 1 is characterised in that, the image-region of cutting apart carefullyyer compared with the past is carried out gray scale respectively handle.
＜constitute
Fig. 1 is the block diagram of the structure of explanation visual processing apparatus 1.Visual processing apparatus 1 has: will be divided into as the original image of input signal IS input a plurality of image-region Pm (1≤m≤n:n is the number of cutting apart of original image) image segmentation portion 2, each image-region Pm is derived the gray-scale transformation curve leading-out portion 10 of gray-scale transformation curve Cm and loads gray-scale transformation curve Cm and the gray scale handling part 5 of the output signal OS that gray scale handles has been carried out in output to each image-region Pm.Gray-scale transformation curve leading-out portion 10 constitutes by the histogram generating unit 3 of lightness (lightness) the histogram Hm of the pixel that generates wide area image-region Em and according to the grey scale curve generating unit 4 that the lightness histogram Hm that is generated generates each image-region Pm corresponding gray scale conversion curve Cm.Wherein, wide area image-region Em is made of the image-region of each image-region Pm and image-region Pm periphery.
＜effect 〉
Use Fig. 2～Fig. 4, the action of each one is described.Image segmentation portion 2 will be divided into a plurality of (n) image-region Pm (with reference to Fig. 2) as the original image of input signal IS input.Here, original image cut apart number, than visual processing apparatus in the past 300 shown in Figure 33 to cut apart number (for example 4～16 cut apart) many, for example, in the horizontal 80 cut apart in the vertical 60 cut apart 4800 cut apart etc.
Histogram generating unit 3 is to the lightness histogram Hm of each image-region Pm generation wide area image-region Em.Here, wide area image-region Em is meant, comprises the set of a plurality of image-regions of each image-region Pm, is vertical 5 of the center, the set of horizontal 5 25 image-regions with image-region Pm for example.In addition, the difference according to the position of image-region Pm can't obtain vertical 5 at the periphery of image-region Pm, horizontal 5 wide area image-region Em sometimes.For example, for the image-region PI of the periphery that is positioned at original image, just can't obtain vertical 5, horizontal 5 wide area image-region EI at the periphery of image-region PI.In this case, will be vertical 5 of the center, horizontal 5 zone and original image overlapping areas with image-region PI, adopt as wide area image-region EI.The lightness histogram Hm that histogram generating unit 3 is generated, the distribution of the brightness value of the both full-pixel in the expression wide area image-region Em.That is, in the lightness histogram Hm shown in Fig. 3 (a)～(c), transverse axis is represented the lightness grade of input signal IS, longitudinal axis remarked pixel number.
Grey scale curve generating unit 4 is according to " pixel count " of the lightness histogram Hm of the order of lightness accumulation wide area image-region Em, with the gray-scale transformation curve Cm (with reference to Fig. 4) of this accumulation curve as image-region Pm.Among the gray-scale transformation curve Cm shown in Fig. 4, transverse axis is represented the brightness value of the pixel of the image-region Pm among the input signal IS, and the longitudinal axis is represented the brightness value of the pixel of the image-region Pm among the output signal OS.Gray scale handling part 5 loads gray-scale transformation curve Cm, and according to gray-scale transformation curve Cm, the brightness value of the pixel of the image-region Pm among the conversion input signal IS.
＜visual processing method and visual processing program 〉
The flow chart of the visual processing method among Fig. 5 in the expression explanation visual processing apparatus 1.Visual processing method shown in Fig. 5 is to realize by hardware in visual processing apparatus 1, and carries out the method for the gray scale processing of input signal IS (with reference to Fig. 1).In the visual processing method shown in Fig. 5, input signal IS is handled (step S10～S16) with image as unit.To be divided into a plurality of image-region Pm (1≤m≤n:n is the number of cutting apart of original image) (step S11) as the original image of input signal IS input, and each image-region Pm be carried out gray scale handle (step S12～S15).
The lightness histogram Fm (step S12) of the pixel of the wide area image-region Em that generation is made of the image-region of each image-region Pm and image-region Pm periphery.And then, according to lightness histogram Hm, generate each image-region Pm corresponding gray scale conversion curve Cm (step S13).Here, omit explanation to lightness histogram Hm and gray-scale transformation curve Cm (with reference to above-mentioned＜effect〉hurdle).Use the gray-scale transformation curve Cm that is generated, the pixel of image-region Pm is carried out gray scale handle (step S14).Afterwards, judge whether to have finished processing (step S15), the processing of step S12～S15 is repeated cutting apart for several times of original image, finish up to judging processing to all images zone Pm.By more than, finish the processing (step S16) of image as unit.
In addition, each step of the visual processing method shown in Fig. 5 can realize as visual processing program by computer etc.
＜effect 〉
(1)
Gray-scale transformation curve Cm generates each image-region Pm.Therefore, compare, can carry out suitable gray scale and handle with the situation of original image all being carried out same greyscale transformation.
(2)
The gray-scale transformation curve Cm that each image-region Pm is generated generates according to the lightness histogram Hm of wide area image-region Em.Therefore, even the size of each image-region Pm is very little, also can obtain enough brightness value sampling.In addition, consequently,, also can generate suitable gray-scale transformation curve Cm even to little image-region Pm.
(3)
The pairing wide area image-region of adjacent image regions has overlapping.Therefore, adjacent image regions corresponding gray scale conversion curve demonstrates similar tendency usually.Therefore, handle the effect of adding spatial manipulation can for the gray scale of each image-region, thereby the border seam that can prevent adjacent image regions become obvious artificially.
(4)
The size of each image-region Pm, compared with the past less.Therefore, can suppress to produce in the image-region Pm false contouring.
＜variation 〉
The present invention is not limited in above-mentioned execution mode, can also carry out various distortion in the scope that does not break away from its main points.
(1)
Though in the above-mentioned execution mode, as original image cut apart the number one of example, adopted 4800 to cut apart, effect of the present invention is not limited in this situation, other cut apart the number in also can obtain same effect.In addition, between the treating capacity and visual effect that gray scale is handled, about cutting apart the relation that there is balance in number.That is, increase if cut apart number, the treating capacity that gray scale is handled also increases, but can access better visual effect (for example the inhibition of false contouring etc.).
(2)
Though in the above-mentioned execution mode, be 25 as one of the number of the image-region that constitutes wide area image-region example, effect of the present invention is not limited in this situation, other number also can obtain same effect.
[the 2nd execution mode]
Contrast Fig. 6～Figure 18 describes the visual processing apparatus 11 as the 2nd execution mode of the present invention.Visual processing apparatus 11 for example is built-in or be connected computer, television set, digital camera, mobile phone, PDA etc. handles in the machine of image, carries out the device of the gray scale processing of image.Visual processing apparatus 11 is characterised in that, switches use and is used as a plurality of gray-scale transformation curves that LUT preserves in advance.
＜constitute
Fig. 6 represents to illustrate the block diagram of the structure of visual processing apparatus 11.Visual processing apparatus 11 has image segmentation portion 12, selects signal leading-out portion 13 and gray scale handling part 20.Image segmentation portion 12, with input signal IS input, and output will be cut apart a plurality of image-region Pm (1≤m≤n:n is the number of cutting apart of original image) that obtain as the original image of input signal IS input.Select signal leading-out portion 13, output is used for selecting the selection signal Sm of the gray-scale transformation curve Cm that the gray scale of each image-region Pm uses in handling.Gray scale handling part 20 has gray scale processing execution portion 14 and gray scale correction portion 15.Gray scale processing execution portion 14, have a plurality of gray-scale transformation curve candidate G1～Gp (p is the candidate number) as two-dimentional LUT, with input signal IS with select signal Sm as input, output has been carried out the gray scale processing signals CS that gray scale is handled to the pixel in each image-region Pm.Gray scale correction portion 15, as input, output has been carried out the output signal OS that revises to the gray scale of gray scale processing signals CS with gray scale processing signals CS.
(about the gray-scale transformation curve candidate)
Use Fig. 7, G1～Gp describes to the gray-scale transformation curve candidate.Gray-scale transformation curve candidate G1～Gp is the curve that provides the relation between the brightness value of pixel of the brightness value of pixel of input signal IS and gray scale processing signals CS.Among Fig. 7, transverse axis is represented the brightness value of the pixel among the input signal IS, and the longitudinal axis is represented the brightness value of the pixel among the gray scale processing signals CS.Gray-scale transformation curve candidate G1～Gp has about the dull relation that reduces of footnote, to the brightness value of the pixel of all input signal IS, satisfy G1 〉=G2 〉=... the relation of 〉=Gp.For example, gray-scale transformation curve candidate G1～Gp is that the brightness value with the pixel of each input signal IS is " power function " of variable, and (1≤m≤p, x are variable under the situation that is expressed as Gm=x^ (δ m), δ m is a constant), satisfy δ 1≤δ 2≤... the relation of≤δ p.Here, the brightness value of input signal IS is in the scope of value [0.0～1.0].
In addition, the relation of above gray-scale transformation curve candidate G1～Gp is in the gray-scale transformation curve candidate bigger for footnote, under the less situation of input signal IS, or the gray-scale transformation curve candidate less for footnote, under the bigger situation of input signal IS, can be false.This is because have this situation hardly, and is very little to the influence of picture element.
Gray scale processing execution portion 14 has gray-scale transformation curve candidate G1～Gp as two-dimentional LUT.Also promptly, two-dimentional LUT is the brightness value and the selection signal Sm that selects gray-scale transformation curve candidate G1～Gp to the pixel of input signal IS, provides the look-up table (LUT) of brightness value of the pixel of gray scale processing signals CS.Illustrated among Fig. 8 one of should two dimension LUT example.Two-dimentional LUT41 shown in Fig. 8 is the matrix of 64 row, 64 row, and each gray-scale transformation curve candidate G1～G64 is arranged on the line direction (laterally).On the matrix column direction (vertically), for example arranging the pixel value of the value corresponding gray scale processing signals CS of the input signal IS that by high 6 value of the pixel value of 10 represented input signal IS, promptly is divided into 64 grades.The pixel value of gray scale processing signals CS is under the situation of " power function " at gray-scale transformation curve candidate G1～Gp, for example has the value of [0.0～1.0] scope.
＜effect 〉
Action to each one describes.The image segmentation portion 2 of image segmentation portion 12 and Fig. 1 almost similarly carries out work, will be divided into a plurality of (n) image-region Pm (with reference to Fig. 2) as the original image of input signal IS input.Here, original image cut apart number, than visual processing apparatus in the past 300 shown in Figure 33 to cut apart number (for example 4～16 cut apart) many, for example in the horizontal 80 cut apart in the vertical 60 cut apart 4800 cut apart etc.
Select signal leading-out portion 13, from gray-scale transformation curve candidate G1～Gp, select the gray-scale transformation curve Cm that uses among each image-region Pm.Specifically, select signal leading-out portion 13, calculate the average brightness value of the wide area image-region Em of image-region Pm,, select any one among gray-scale transformation curve candidate G1～Gp according to the average brightness value that calculates.That is, gray-scale transformation curve candidate G1～Gp is associated with the average brightness value of wide area image-region Em, and average brightness value is big more, selects the big more gray-scale transformation curve candidate G1～Gp of footnote.
Here, so-called wide area image-region Em and uses illustrated identical of Fig. 2 in [the 1st execution mode].That is, wide area image-region Em is the set that comprises a plurality of image-regions of each image-region Pm, is vertical 5 of the center, the set of horizontal 5 25 image-regions with image-region Pm for example.In addition, according to the difference of the position of image-region Pm, can't obtain vertical 5, horizontal 5 wide area image-region Em at the periphery of image-region Pm sometimes.For example, for the image-region PI of the periphery that is positioned at original image, just can't obtain vertical 5, horizontal 5 wide area image-region EI at the periphery of image-region PI.In this case, will be vertical 5 of the center, horizontal 5 zone and original image overlapping areas with image-region PI, adopt as wide area image-region EI.
Select the selection result of signal leading-out portion 13, as any one the selection signal Sm output among expression gray-scale transformation curve candidate G1～Gp.More specifically, select signal Sm, be used as footnote (1～p) the value output of gray-scale transformation curve candidate G1～Gp.
Gray scale processing execution portion 14, the brightness value of the pixel of the image-region Pm that input signal IS is comprised and select signal Sm as input for example, uses the two-dimentional LUT41 shown in Fig. 8, the brightness value of output gray level processing signals CS.
Gray scale correction portion 15, according to locations of pixels and gray-scale transformation curve that the ambient image regions of image-region Pm and image-region Pm is selected, the brightness value of the pixel of the image-region PM that gray scale processing signals CS is comprised is revised.For example, with the interior proportion by subtraction of location of pixels, the gray-scale transformation curve of revising the gray-scale transformation curve Cm that is applicable to the pixel that image-region Pm comprises and the image-region of the periphery of image-region Pm being selected, and obtain the brightness value of revised pixel.
Contrast Fig. 9 is described in more details the action of gray scale correction portion 15.What represent among Fig. 9 is, gray-scale transformation curve Co, Cp, Cq, the Cr of image-region Po, Pp, Pq, Pr (o, p, q, r are cut apart the following positive integer of several n (with reference to Fig. 2)) are selected as gray-scale transformation curve candidate Gs, Gt, Gu, Gv (s, t, u, v are that the candidate of gray-scale transformation curve is counted the following positive integer of p).
Here, if position as the pixel x (being made as brightness value [x]) of the image-region Po of the object of gray scale correction, be arranged in the center of image-region Po and intracardiac being divided into [i:1-i] of image-region Pp, and on the position with intracardiac being divided into [j:1-j] among the center of image-region Po and the image-region Pq.In this case, the brightness value of the revised pixel x of gray scale [x '] is obtained and is [x ']={ (1-j) (1-i) [Gs]+(1-j) (i) [Gt]+(j) (1-i) [Gu]+(j) (i) [Gv] } { [x]/[Gs] }.In addition, [Gs], [Gt], [Gu], [Gv] are to the brightness value under the situation of brightness value [x] application gray-scale transformation curve candidate Gs, Gt, Gu, Gv.
＜visual processing method and visual processing program 〉
The flow chart of the visual processing method among Figure 10 in the expression explanation visual processing apparatus 11.Visual processing method shown in Figure 10 is to realize by hardware in visual processing apparatus 11, and carries out the method for the gray scale processing of input signal IS (with reference to Fig. 6).In the visual processing method shown in Figure 10, input signal IS is handled (step S20～S26) with image as unit.To be divided into a plurality of image-region Pm (1≤m≤n:n is the number of cutting apart of original image) (step S21) as the original image of input signal IS input, and each image-region Pm be carried out gray scale handle (step S22～S24).
In the processing of each image-region Pm, from gray-scale transformation curve candidate G1～Gp, select the gray-scale transformation curve Cm (step S22) that uses among each image-region Pm.Specifically, calculate the average brightness value of the wide area image-region Em of image-region Pm,, select any one among gray-scale transformation curve candidate G1～Gp according to the average brightness value that calculates.Gray-scale transformation curve candidate G1～Gp is associated with the average brightness value of wide area image-region Em, and average brightness value is big more, selects the big more gray-scale transformation curve candidate G1～Gp of footnote.The explanation of omission wide area image-region Em here, (with reference to above-mentioned＜effect〉hurdle).
Among the brightness value of the pixel of the image-region Pm that input signal IS is comprised and the expression gray-scale transformation curve candidate G1～Gp by the selection signal Sm of the selected gray-scale transformation curve candidate of step S22, for example use the two-dimentional LUT41 shown in Fig. 8, the brightness value (step S23) of output gray level processing signals CS.Afterwards, judge whether to have finished processing (step S24), the processing of step S22～S24 is repeated cutting apart for several times of original image, finish up to judgment processing to all images zone Pm.By more than, finish the processing of image-region unit.
The brightness value of the pixel of the image-region Pm that gray scale processing signals CS is comprised is by according to locations of pixels and gray-scale transformation curve correction (step S25) that the image-region of image-region Pm and image-region Pm periphery is selected.For example, with the interior proportion by subtraction of location of pixels, the gray-scale transformation curve of revising the gray-scale transformation curve Cm that is used for the pixel that image-region Pm comprised and the image-region of the periphery of image-region Pm being selected, and obtain the brightness value of revised pixel.About the detailed content of revising, and omission explanation here (with reference to above-mentioned＜effect〉hurdle, Fig. 9).
By like this, finish the processing (step S26) of image as unit.
In addition, each step of the visual processing method shown in Figure 10 can realize as visual processing program by computer etc.
＜effect 〉
By the present invention, can access with above-mentioned [the 1st execution mode]＜effect much at one effect.Below, the distinctive effect of the 2nd execution mode is described.
(1)
To the gray-scale transformation curve Cm of each image-region Pm selection, according to the average brightness value generation of wide area image-region Em.Therefore, even image-region Pm's is big or small less, also can obtain the sampling of enough brightness values.In addition, consequently,, also can select suitable gray-scale transformation curve Cm to use for less image-region Pm.
(2)
Gray scale processing execution portion 14, the two-dimentional LUT that has in advance to be generated.Therefore, can cut down gray scale and handle needed processing load, in particular, can cut down the needed processing load of generation of gray-scale transformation curve Cm.Consequently, can allow the gray scale of image-region Pm handle needed processing high speed.
(3)
Gray scale processing execution portion 14 uses two-dimentional LUT to carry out gray scale and handles.Two dimension LUT is read from storage devices such as hard disk that visual processing apparatus 11 had or ROM and is used for gray scale and handled.By changing the content of the two-dimentional LUT that is read, do not need to change the formation of hardware, just can realize various gray scales processing.That is, can realize suitable gray scale processing according to the characteristic of original image.
(4)
Gray scale correction portion 15 is revised the gray scale of the pixel of implementing the image-region Pm that gray scale handles with 1 gray-scale transformation curve Cm.Therefore, can access the output signal OS that has implemented more suitable gray scale processing.For example, can suppress the generation of false contouring.In addition, can prevent further that among the output signal OS, the seam on the border of each image-region Pm becomes eye-catching artificially.
＜variation 〉
The present invention is not limited in above-mentioned execution mode, can also carry out various distortion in the scope that does not break away from its main points.
(1)
Though in the above-mentioned execution mode, as original image cut apart the number one of example, adopted 4800 to cut apart, effect of the present invention is not limited in this situation, adopt other cut apart the number also can obtain same effect.In addition, between the treating capacity and visual effect that gray scale is handled, about cutting apart the relation that there is balance in number.That is, increase if cut apart number, the treating capacity that gray scale is handled also increases, but can access better visual effect (for example having suppressed the image of false contouring etc.).
(2)
Though in the above-mentioned execution mode, be 25 as one of the number of the image-region that constitutes wide area image-region example, effect of the present invention is not limited in this situation, adopts other number also can obtain same effect.
(3)
In the above-mentioned execution mode, the two-dimentional LUT41 that 64 row, 64 column matrix are constituted is as one of two-dimentional LUT example.Here, effect of the present invention is not limited in the two-dimentional LUT of this size.For example, can also be the matrix that is arranged with more gray-scale transformation curve candidates on the line direction.In addition, can also be on the matrix column direction, arrange the pixel value of the pixel value of input signal IS being divided into the value corresponding gray scale processing signals CS behind the thinner stride.Specifically, for example can arrange the pixel value of gray scale processing signals CS to each pixel value by the input signal IS of 10 bit representations.
If the size of two-dimentional LUT is bigger, then can carry out more suitable gray scale and handle, if less, then can cut down the memory of preserving two-dimentional LUT.
(4)
What illustrate in the above-mentioned execution mode is, on the matrix column direction, for example is arranged with high 6 value by the pixel value of the input signal IS of 10 bit representations, promptly has been divided into the pixel value of value corresponding gray scale processing signals CS of 64 grades input signal IS.Here, gray scale processing signals CS also can be used as by gray scale processing execution portion 14, the composition output of the matrix behind low 4 the value linear interpolation of the pixel value of usefulness input signal IS.Promptly, on the matrix column direction, for example be arranged with composition by high 6 the pairing matrix of value of the pixel value of the input signal IS of 10 bit representations, low 4 value with the pixel value of input signal IS, to the composition of high 6 the pairing matrix of value of the pixel value of input signal IS with give composition that high 6 value of the pixel value of input signal IS adds the pairing matrix of value after [1] (for example, among Fig. 8, the composition under 1 row) carry out linear interpolation, and export as gray scale processing signals CS.
By like this,, also can carry out more suitable gray scale and handle even two-dimentional LUT41's (with reference to Fig. 8) is big or small less.
(5)
What illustrate in the above-mentioned execution mode is according to the average brightness value of wide area image-region EM, to select the gray-scale transformation curve Cm that uses among the image-region Pm.Here, the system of selection of gray-scale transformation curve Cm is not limited in this method.For example, can also select the gray-scale transformation curve Cm that uses among the image-region Pm according to maximum brightness value or the minimum brightness value of wide area image-region Em.In addition, when selecting gray-scale transformation curve Cm, selecting the value [Sm] of signal Sm, can be the average brightness value of wide area image-region Em, maximum brightness value or minimum brightness value itself.In this case, the value that will select signal Sm to be got is divided into 64 grades of each values that obtain, related gray-scale transformation curve candidate G1～G64.
In addition, the gray-scale transformation curve Cm that uses in for example can following selection image-region Pm.That is, Pm obtains average brightness value to each image-region, according to each average brightness value, obtains the interim selection signal Sm ' of each image-region Pm correspondence.Here, the value of interim selection signal Sm ' is the numbering of the footnote of gray-scale transformation curve candidate G1～Gp.And then, each image-region that wide area image-region Em is comprised, value equalization with interim selection signal Sm ', obtain the value [Sm] of the selection signal Sm of image-region Pm, and be the candidate of footnote with the integer with closest value [Sm] among gray-scale transformation curve candidate G1～Gp, be chosen as gray-scale transformation curve Cm.
(6)
What illustrate in the above-mentioned execution mode is according to the average brightness value of wide area image-region Em, to select the gray-scale transformation curve Cm that uses among the image-region Pm.Here, can be not yet simple average according to wide area image-region Em, but, select the gray-scale transformation curve Cm that uses among the image-region Pm according to weighted average (having the average of weight).For example shown in Figure 11, obtain the average brightness value of each image-region that constitutes wide area image-region Em, image-region Ps1, the Ps2 of the average brightness value that the average brightness value that has with image-region Pm is differed greatly ... alleviate its weight, or it is removed, obtain the average brightness value of wide area image-region Em.
By like this, even (for example comprise at wide area image-region Em under the situation of lightness special region, wide area image-region Em comprises under the situation on border of the different object of two brightness values), for the selection of the gray-scale transformation curve Cm that uses among the image-region Pm, the influence that brightness value brought of this special area is also less, handles thereby can carry out more suitable gray scale.
(7)
In the above-mentioned execution mode, the existence of gray scale correction portion 15 is arbitrarily.Promptly, even under the situation of output gray level processing signals CS, compare with visual processing apparatus 300 (with reference to Figure 33) in the past, also can access with [the 1st execution mode]＜effect in the identical effect of content put down in writing, and with [the 2nd execution mode]＜effect the identical effect of content put down in writing in (1) and (2).
(8)
In the above-mentioned execution mode, gray-scale transformation curve candidate G1～Gp has about the dull relation that reduces of footnote, to the brightness value of the pixel of all input signal IS, satisfy G1 〉=G2 〉=... the relation of 〉=Gp.Here, have gray-scale transformation curve candidate G1～Gp of two-dimentional LUT, also can be to the part of the brightness value of the pixel of input signal IS, do not satisfy G1 〉=G2 〉=... the relation of 〉=Gp.Also promptly, any one of gray-scale transformation curve candidate G1～Gp can be relation intersected with each other.
For example, be in little light in the darker night scene etc. (be in the night scene neon light part etc.), though the value of input signal IS is bigger, but the average brightness value of wide area image-region Em is less, in this case, it is very little to the influence that picture element brings to implement the value of the picture signal that gray scale handles.In this case, have gray-scale transformation curve candidate G1～Gp of two-dimentional LUT, can be to the part of the brightness value of the pixel of input signal IS, do not satisfy G1 〉=G2 〉=... the relation of 〉=Gp.Also promptly, in the less part of the influence that the value after gray scale is handled is brought to picture element, the value that two-dimentional LUT preserved can be arbitrarily.
In addition,, preferably also allow to the input signal IS of same value and the value of selecting signal Sm to be preserved under the situation arbitrarily in value that two-dimentional LUT preserved, keeping relative input signal IS increases or the dull relation that reduces with the value of selecting signal Sm is dull.
In addition, in the above-mentioned execution mode, be that the situation of " power function " is illustrated to gray-scale transformation curve candidate G1～Gp with two-dimentional LUT.Here, gray-scale transformation curve candidate G1～Gp can be formulated as " power function " imprecisely.It can also be function with shapes such as S word, anti-S words.
(9)
In the visual processing apparatus 11, can also have the data of description generating unit of generation as the data of description (profile data) of the value that two-dimentional LUT preserved.Specifically, the data of description generating unit is made of the image segmentation portion 2 in the visual processing apparatus 1 (with reference to Fig. 1) and gray-scale transformation curve leading-out portion 10, and the set with a plurality of gray-scale transformation curves of being generated is kept among the two-dimentional LUT as data of description.
In addition, each gray-scale transformation curve of being preserved among the two-dimentional LUT can also be associated with the input signal IS that has implemented spatial manipulation.In this case, in the visual processing apparatus 11, image segmentation portion 12 and selection signal leading-out portion 13 can be replaced with the spatial manipulation portion that input signal IS is carried out spatial manipulation.
(10)
In the above-mentioned execution mode, the brightness value of the pixel of input signal IS can not be the value in [0.0～1.0] scope.Under the situation of input signal IS as the value input of other scopes, can be with value value of being standardized as [0.0～1.0] of this scope.In addition, also can not carry out standardization, and handled value in the above-mentioned processing is suitably changed.
(11)
Each gray-scale transformation curve candidate G1～Gp can be the input signal IS with dynamic range bigger than common dynamic range to be implemented gray scale handle, and export the gray-scale transformation curve of the gray scale processing signals CS of common dynamic range.
In recent years, by concentrate light quantity use S/N preferably CCD, the methods such as transducer with the highly sensitive pixel of muting sensitivity are opened or used to electronic shutter length for twice, the exploitation of machine that can handle the dynamic range bigger 1～3 than common dynamic range is constantly developed.
Meanwhile, even need to have under the situation of the dynamic range bigger, also can carry out suitable gray scale and handle than common dynamic range (for example signal of the scope of value [0.0～1.0]) at input signal IS.
Here, as shown in figure 12, the gray-scale transformation curve of use, to having exceeded the range input signal IS of value [0.0～1.0], the also gray scale processing signals CS of output valve [0.0～1.0].
By like this,, also can carry out suitable gray scale and handle, and export the gray scale processing signals CS of common dynamic range for having than for the input signal IS of wide dynamic range.
In addition, in the above-mentioned execution mode, " pixel value of gray scale processing signals CS is under the situation of " power function " at gray-scale transformation curve candidate G1～Gp, for example has the value of [0.0～1.0] scope ".Here, the pixel value of gray scale processing signals CS is not limited in this scope.For example, for the input signal IS of value [0.0～1.0], gray-scale transformation curve candidate G1～Gp can carry out dynamic range compression.
(12)
In the above-mentioned execution mode, " gray scale processing execution portion 14 has gray-scale transformation curve candidate G1～Gp as two-dimentional LUT " is illustrated.Here, gray scale processing execution portion 14 also can have one dimension LUT, and storage is used for determining the parameter of curve of gray-scale transformation curve candidate G1～Gp and the relation between the selection signal Sm.
" formation "
The expression explanation is as the block diagram of the structure of the gray scale processing execution portion 44 of the variation of gray scale processing execution portion 14 among Figure 13.Gray scale processing execution portion 44, with input signal IS with select signal Sm as input, output is gray scale processing signals CS as the input signal IS that was handled by gray scale.Gray scale processing execution portion 44 has parameter of curve efferent 45 and operational part 48.
Parameter of curve efferent 45 is made of 1LUT46 and 2LUT47.1LUT46 and 2LUT47 will select signal Sm as input, and parameter of curve P1 and the P2 of the specified gray-scale transformation curve candidate Gm of signal Sm selected in output respectively.
" about one dimension LUT "
1LUT46 and 2LUT47 are the one dimension LUT that preserves the value of selecting pairing parameter of curve P1 of signal Sm and P2 respectively.Before 1LUT46 and 2LUT47 are elaborated, the content of parameter of curve P1 and P2 is described.
Contrast Figure 14, to parameter of curve P1 and P2, and gray-scale transformation curve candidate G1～Gp between relation describe.Figure 14 represents gray-scale transformation curve candidate G1～Gp.Here, gray-scale transformation curve candidate G1～Gp has about the dull relation that reduces of footnote, to the brightness value of the pixel of all input signal IS, satisfy G1 〉=G2 〉=... the relation of 〉=Gp.In addition, the relation of above gray-scale transformation curve candidate G1～Gp, for the bigger gray-scale transformation curve candidate of footnote under the less situation of input signal IS, perhaps, inferior for the gray-scale transformation curve candidate that footnote is less in the bigger situation of input signal IS, can be false.
Parameter of curve P1 and P2, the value that is used as the set-point corresponding gray scale processing signals CS of input signal IS is exported.Promptly, specifying under the situation of gray-scale transformation curve candidate Gm by selection signal Sm, the value of parameter of curve P1, be used as value [R1m] output of set-point [X1] the corresponding gray scale conversion curve candidate Gm of input signal IS, the value of parameter of curve P2, the value [R2m] that is used as set-point [X2] the corresponding gray scale conversion curve candidate Gm of input signal IS is exported.Here, value [X2] is the value greater than value [X1].
Next, 1LUT46 and 2LUT47 are described.
1LUT46 and 2LUT47 preserve the value of selecting pairing parameter of curve P1 of signal Sm and P2 respectively.In particular, for example for each selection signal Sm that the signal that is used as 6 provides, the value of parameter of curve P1 and P2 provides by 6 respectively.Here, the figure place to selecting signal Sm and parameter of curve P1 and P2 to guarantee is not limited in this.
Use Figure 15, to parameter of curve P1 and P2, describe with relation between the selection signal Sm.The variation of selecting the value of pairing parameter of curve P1 of signal Sm and P2 has been shown among Figure 15.Among 1LUT46 and the 2LUT47, preserve the value that each selects pairing parameter of curve P1 of signal Sm and P2.For example, as the value of selecting the pairing parameter of curve P1 of signal Sm, the value of preserving [R1m], as the value of parameter of curve P2, the value of preserving [R2m].
By above-mentioned 1LUT46 and 2LUT47, to the selection signal Sm that is imported, curve of output parameter P1 and P2.
" about operational part 48 "
At first, operational part 48 with the value of input signal IS, compares with specified value [X1], [X2].
Value (value of being made as [X]) at input signal IS is more than [0.0] and under the situation of discontented [X1], on the initial point in connecting Figure 14 and the straight line of coordinate ([X1], [R1m]), and the value (value of being made as [Y]) of the value of obtaining [X] corresponding gray scale processing signals CS.More specifically, value [Y] is by following formula, i.e. [Y]=([X]/[X1]) * [R1m] obtains.
Value at input signal IS is more than [X1] and under the situation of discontented [X2], on the coordinate ([X1], [R1m]) in connecting Figure 14 and the straight line of coordinate ([X2], [R2m]), and the pairing value of the value of obtaining [X] [Y].More specifically, value [Y] is by following formula, i.e. [Y]=[R1m]+{ ([R2m]-[R1m])/([X2]-[X1]) } * ([X]-[X1]) obtains.
In the value of input signal IS is more than [X2] under the situation below [1.0], on the coordinate ([X2], [R2m]) in connecting Figure 14 and the straight line of coordinate ([1.0], [1.0]), and the pairing value of the value of obtaining [X] [Y].More specifically, value [Y] is by following formula, i.e. [Y]=[R2m]+{ ([1.0]-[R2m])/([1.0]-[X2]) } * ([X]-[X2]) obtains.
By above-mentioned computing, operational part 48 is derived input signal IS corresponding gray scale processing signals CS.
" gray scale processing method program "
Above-mentioned processing can be used as the gray scale handling procedure, by execution such as computers.The gray scale handling procedure is to be used for allowing computer carry out the program of the gray scale processing method of the following stated.
Gray scale processing method is to obtain input signal IS and select signal Sm, and the method for output gray level processing signals CS, it is characterized in that: use one dimension LUT that input signal IS is carried out gray scale and handle.
At first, obtain and select after the signal Sm, from 1LUT46 and 2LUT47, curve of output parameter P1 and P2.Here, omission is to the detailed description of 1LUT46,2LUT47, parameter of curve P1 and P2.
And then, according to parameter of curve P1 and P2, carry out the gray scale of input signal IS and handle.Detailed content about gray scale is handled is described in the explanation to operational part 48, therefore explanation here.
By above-mentioned gray scale processing method, derive input signal IS corresponding gray scale processing signals CS.
" effect "
Gray scale processing execution portion 44 as the variation of gray scale processing execution portion 14 has two one dimension LUT rather than two-dimentional LUT.Therefore, can cut down the memory capacity that is used for preserving look-up table.
" variation "
(1)
In the above-mentioned variation, " value of parameter of curve P1 and P2 is the value of the set-point corresponding gray scale conversion curve candidate Gm of input signal IS " is illustrated.Here, parameter of curve P1 and P2 also can be other the parameters of curve of gray-scale transformation curve candidate Gm.Below be specifically described.
(1-1)
Parameter of curve can be the slope of gray-scale transformation curve candidate Gm.Contrast Figure 14 is specifically described.Specifying under the situation of gray-scale transformation curve candidate Gm with selection signal Sm, the value of parameter of curve P1, it is the value [K1m] of the slope of the gray-scale transformation curve candidate Gm in the given range [0.0～X1] of input signal IS, the value of parameter of curve P2 is the value [K2m] of the slope of the gray-scale transformation curve candidate Gm in the given range [X1～X2] of input signal IS.
Use Figure 16, to parameter of curve P1 and P2, describe with relation between the selection signal Sm.What represent among Figure 16 is to select the variation of the value of pairing parameter of curve P1 of signal Sm and P2.Among 1LUT46 and the 2LUT47, preserve the value that each selects pairing parameter of curve P1 of signal Sm and P2.For example, as value value of preserving [K1m] of selecting the pairing parameter of curve P1 of signal Sm, as value value of preserving [K2m] of parameter of curve P2.
By above-mentioned 1LUT46 and 2LUT47, to the selection signal Sm that is imported, curve of output parameter P1 and P2.
In the operational part 48,, derive input signal IS corresponding gray scale processing signals CS according to obtained parameter of curve P1 and P2.Concrete order is as described below.
At first, operational part 48 is compared the value of input signal IS with specified value [X1], [X2].
Value (value of being made as [X]) at input signal IS is more than [0.0] and under the situation of discontented [X1], initial point and coordinate ([X1] in connecting Figure 14, [K1m] * [X1] (below be called " Y1 ")) on the straight line, the value (value of being made as [Y]) of the value of obtaining [X] corresponding gray scale processing signals CS.More specifically, value [Y] is by following formula, i.e. [Y]=[K1m] * [X] obtains.
Value at input signal IS is more than [X1] and under the situation of discontented [X2], coordinate ([X1] in connecting Figure 14, [Y1]) with the straight line of coordinate ([X2], [K1m] * [X1]+[K2m] * ([X2]-[X1]) (below be called " Y2 ")) on, the pairing value of the value of obtaining [X] [Y].More specifically, value [Y] is by following formula, i.e. [Y]=[Y1]+[K2m] * ([X]-[X1]) obtains.
Value at input signal IS is more than [X2] and under the situation below [1.0], on the coordinate ([X2], [Y2]) in connecting Figure 14 and the straight line of coordinate ([1.0], [1.0]), and the pairing value of the value of obtaining [X] [Y].More specifically, value [Y] is by following formula, i.e. [Y]=[Y2]+{ ([1.0]-[Y2])/([1.0]-[X2]) } * ([X]-[X2]) obtains.
By above-mentioned computing, operational part 48 is derived input signal IS corresponding gray scale processing signals CS.
(1-2)
Parameter of curve can also be the coordinate on the gray-scale transformation curve candidate Gm.Contrast Figure 17 is specifically described.Specifying under the situation of gray-scale transformation curve candidate Gm by selection signal Sm, the value of parameter of curve P1, be the value [Mm] of a composition of the coordinate on the gray-scale transformation curve candidate Gm, the value of parameter of curve P2 is the value [Nm] of another composition of the coordinate on the gray-scale transformation curve candidate Gm.In addition, gray-scale transformation curve candidate G1～Gp all is by coordinate (X1, curve Y1).
Use Figure 18, to parameter of curve P1 and P2, describe with relation between the selection signal Sm.What represent among Figure 18 is to select the variation of the value of pairing parameter of curve P1 of signal Sm and P2.Among 1LUT46 and the 2LUT47, preserve the value that each selects pairing parameter of curve P1 of signal Sm and P2.For example, as the value of selecting the pairing parameter of curve P1 of signal Sm, the value of preserving [Mm], as the value of parameter of curve P2, the value of preserving [Nm].
By above-mentioned 1LUT46 and 2LUT47, to the selection signal Sm that is imported, curve of output parameter P1 and P2.
In the operational part 48, by with contrast Figure 14 illustrated identical processing of variation, derive gray scale processing signals CS according to input signal IS.Here omit detailed explanation.
(1-3)
Above variation only is an example, and parameter of curve P1 and P2 can also be other the parameters of curve of gray-scale transformation curve candidate Gm.
In addition, the number of parameter of curve also is not limited only to above-mentioned.Can be still less or more.
In the explanation to operational part 48, be that the computing under the situation of the curve that constitutes of the line segment by straight line is illustrated to gray-scale transformation curve candidate G1～Gp.Here, be used as under the situation that parameter of curve provides, can generate smoothed curve (curve fit), and use the curve that is generated, carry out greyscale transform process by given coordinate at the coordinate on gray-scale transformation curve candidate G1～Gp.
(2)
In the above-mentioned variation, " parameter of curve efferent 45 is made of 1LUT46 and 2LUT47 " is illustrated.Here, parameter of curve efferent 45, the LUT that also can not have the value of the pairing parameter of curve P1 of value that preserve to select signal Sm and P2.
In this case, parameter of curve efferent 45 calculates the value of parameter of curve P1 and P2.In particular, parameter of curve efferent 45, the parameter of preserving the curve chart of parameter of curve P1 shown in expression Figure 15, Figure 16, Figure 18 etc. and P2.Parameter of curve efferent 45 is determined the curve chart of parameter of curve P1 and P2 according to the parameter of being preserved.In addition, use the curve chart of parameter of curve P1 and P2, export the value of selecting pairing parameter of curve P1 of signal Sm and P2.
Here, the parameter that is used for determining the curve chart of parameter of curve P1 and P2 is meant the coordinate on the curve chart, the slope of curve chart, curvature etc.For example, parameter of curve efferent 45 is preserved each 2 the coordinate on the curve of parameter of curve P1 shown in Figure 15 and P2, will connect the straight line of this coordinate of 2, uses as the curve chart of parameter of curve P1 and P2.
Here, when determining the curve chart of parameter of curve P1 and P2, not only can use straight line approximate, can also use piecewise linear approximation, curve approximation etc. according to parameter.
By like this, do not use the memory that is used for preserving LUT, also can the curve of output parameter.Also promptly, the further capacity of the memory that cutting device had.
[the 3rd execution mode]
Contrast Figure 19～Figure 21 describes the visual processing apparatus 21 as the 3rd execution mode of the present invention.Visual processing apparatus 21 for example is built-in or be connected computer, television set, digital camera, mobile phone, PDA etc. handles in the machine of image, carries out the device of the gray scale processing of image.Visual processing apparatus 21 is characterised in that, each is become the object pixels that gray scale is handled, and switches to use a plurality of gray-scale transformation curves of preserving in advance as LUT.
＜constitute
Figure 19 is the block diagram of the structure of explanation visual processing apparatus 21.Visual processing apparatus 21 has image segmentation portion 22, selects signal leading-out portion 23 and gray scale handling part 30.Image segmentation portion 22, as input, output will be cut apart a plurality of image-region Pm (1≤m≤n:n is the number of cutting apart of original image) that obtain as the original image of input signal IS input with input signal IS.Select signal leading-out portion 23, output is used for to the selection signal Sm of each image-region Pm selection gray-scale transformation curve Cm.Gray scale handling part 30 has the signal correction portion 24 of selection and gray scale processing execution portion 25.Select signal correction portion 24, will select signal Sm as input, the signal that output obtains after the selection signal Sm enforcement of each image-region Pm is revised is the selection signal SS of each pixel.Gray scale processing execution portion 25, have a plurality of gray-scale transformation curve candidate G1～Gp (p is the candidate number) as two-dimentional LUT, as input, the output signal OS that obtains after gray scale is handled is implemented in output to each pixel with the selection signal SS of input signal IS and each pixel.
(about the gray-scale transformation curve candidate)
About gray-scale transformation curve candidate G1～Gp and since with " the 2nd execution mode " in use illustrated about the same of Fig. 7, therefore omit explanation here.But in the present embodiment, gray-scale transformation curve candidate G1～Gp is the curve that provides the relation between the brightness value of pixel of the brightness value of pixel of input signal IS and output signal OS.
Gray scale processing execution portion 25 has gray-scale transformation curve candidate G1～Gp as two-dimentional LUT.Also promptly, two-dimentional LUT is to the brightness value of the pixel of input signal IS and select the selection signal SS of gray-scale transformation curve candidate G1～Gp, provides the look-up table (LUT) of brightness value of the pixel of output signal OS.Object lesson is much at one illustrated with " the 2nd execution mode " middle Fig. 8 that uses, and therefore omits explanation here.But in the present embodiment, on the matrix column direction, for example be arranged with pixel value by high 6 the pairing output signal OS of value of the pixel value of the input signal IS of 10 bit representations.
＜effect 〉
Action to each one describes.Image segmentation portion 22 and the image segmentation portion 2 of Fig. 1 almost similarly carry out work, will be divided into a plurality of (n) image-region Pm (with reference to Fig. 2) as the original image of input signal IS input.Here, original image cut apart number, than visual processing apparatus in the past 300 shown in Figure 33 to cut apart number (for example 4～16 cut apart) many, for example for transversely 80 cut apart vertically last 60 cut apart 4800 cut apart.
Select signal leading-out portion 23, each image-region Pm is selected gray-scale transformation curve Cm from gray-scale transformation curve candidate G1～Gp.Specifically, select signal leading-out portion 23, calculate the average brightness value of the wide area image-region Em of image-region Pm,, select any one among gray-scale transformation curve candidate G1～Gp according to the average brightness value that calculates.Also promptly, gray-scale transformation curve candidate G1～Gp is associated with the average brightness value of wide area image-region Em, and average brightness value is big more, selects the big more gray-scale transformation curve candidate G1～Gp of footnote.
Here, wide area image-region Em and uses illustrated identical of Fig. 2 in [the 1st execution mode].Also promptly, wide area image-region Em is the set that comprises a plurality of image-regions of each image-region Pm, for example is to be the set of 25 image-regions of vertical 5 horizontal 5 at center with image-region Pm.In addition, the difference according to the position of image-region Pm can't obtain vertical 5 horizontal 5 wide area image-region Em at the periphery of image-region Pm sometimes.For example, for the image-region PI of the periphery that is positioned at original image, just can't obtain vertical 5 horizontal 5 wide area image-region EI at the periphery of image-region PI.In this case, will be vertical 5 horizontal 5 the zones and the original image overlapping areas at center with image-region PI, adopt as wide area image-region EI.
Select the selection result of signal leading-out portion 23, be used as any one the selection signal Sm output among expression gray-scale transformation curve candidate G1～Gp.More specifically, select signal Sm, be used as footnote (1～p) the value output of gray-scale transformation curve candidate G1～Gp.
Select signal correction portion 24, implement to revise by using each selection signal Sm to each image-region Pm output, output is used for to the selection signal SS of each pixel of each the pixel selection gray-scale transformation curve that constitutes input signal IS.For example, the pairing selection signal of the pixel that comprises among image-region Pm SS is used the interior proportion by subtraction correction of location of pixels that the value of the selection signal of the image-region output of the periphery of image-region Pm and image-region Pm is obtained.
Contrast Figure 20 is described in more details the action of selecting signal correction portion 24.What represent among Figure 20 is that the state of signal So, Sp, Sq, Sr is selected in image-region Po, Pp, Pq, Pr (o, p, q, r are cut apart the following positive integer of several n (with reference to Fig. 2)) output.
Here, establish position as the object pixels x of gray scale correction, for intracardiac being divided into [i:1-i] among the center of image-region Po and the image-region Pp, and with the position of intracardiac being divided into [j:1-j] among the center of image-region Po and the image-region Pq.In this case, the value [SS] of the pairing selection signal of pixel x SS is obtained by [SS]={ (1-j) (1-i) [So]+(1-j) (i) [Sp]+(j) (1-i) [Sq]+(j) (i) [Sr] }.In addition, [So], [Sp], [Sq], [Sr] are the values of selecting signal So, Sp, Sq, Sr.
Gray scale processing execution portion 25, the brightness value of the pixel that input signal IS is comprised and select signal SS as input uses for example two-dimentional LUT41 shown in Fig. 8, with the brightness value output of output signal OS.
In addition, in the value [SS] of selecting signal SS, the footnote that is not gray-scale transformation curve candidate G1～Gp of being had with two-dimentional LUT41 is (under the situation of 1～p) value that equates, in the gray scale of input signal IS is handled, use with the integer of closest value [SS] gray-scale transformation curve candidate G1～Gp as footnote.
＜visual processing method and visual processing program 〉
The flow chart of the visual processing method among Figure 21 in the expression explanation visual processing apparatus 21.Visual processing method shown in Figure 21 is to realize by hardware in visual processing apparatus 21, and carries out the method for the gray scale processing of input signal IS (with reference to Figure 19).In the visual processing method shown in Figure 21, input signal IS is handled (step S30～S37) with image as unit.Will be as the original image of input signal IS input, be divided into a plurality of image-region Pm (1≤m≤n:n is the number of cutting apart of original image) (step S31), each image-region Pm is selected gray-scale transformation curve Cm (step S32～S33), according to the selection signal Sm that is used for to each image-region Pm selection gray-scale transformation curve Cm, to each pixel selection gray-scale transformation curve of original image, and carry out gray scale under the pixel unit and handle (step S34～S36).
Below, each step is specifically described.
From gray-scale transformation curve candidate G1～Gp, each image-region Pm is selected gray-scale transformation curve Cm (step S32).Specifically, calculate the average brightness value of the wide area image-region Em of image-region Pm,, select any one among gray-scale transformation curve candidate G1～Gp according to the average brightness value that calculates.Gray-scale transformation curve candidate G1～Gp is associated with the average brightness value of wide area image-region Em, and average brightness value is big more, selects the big more gray-scale transformation curve candidate G1～Gp of footnote.Here omit wide area image-region Em explanation (with reference to above-mentioned＜effect hurdle).The result who selects is output as any one the selection signal Sm of expression gray-scale transformation curve candidate G1～Gp.More specifically, select signal Sm, be used as footnote (1～p) the value output of gray-scale transformation curve candidate G1～Gp.Afterwards, judge whether to have finished processing (step S33), the processing of step S32～S33 is repeated cutting apart for several times of original image, finish up to judging processing to all images zone Pm.By more than, finish the processing of image-region unit.
Select signal Sm to implement to revise by using to each of each image-region Pm output, export the selection signal SS (step S34) that is used for to each pixel of each pixel selection gray-scale transformation curve of constituting input signal IS.For example, the pairing selection signal of the pixel that comprises among image-region Pm SS is used the interior proportion by subtraction of location of pixels, revises the value of the selection signal of the image-region output of the periphery of image-region Pm and image-region Pm is obtained.Omission about the detailed content revised (with reference to above-mentioned＜effect〉hurdle, Figure 20).
The brightness value of the pixel that input signal IS is comprised and select signal SS as input for example uses the two-dimentional LUT41 shown in Fig. 8, with the brightness value output (step S35) of output signal OS.Afterwards, judge whether to have finished the processing (step S36) to all pixels, the processing repetition pixel several with step S34～S36 finishes up to judging processing.By more than, finish the processing of image as unit.
In addition, each step of the visual processing method shown in Figure 21 can realize as visual processing program by computer etc.
＜effect 〉
By the present invention, can access with above-mentioned [the 1st execution mode] and [the 2nd execution mode]＜effect much at one effect.Below, the distinctive effect of the 3rd execution mode is described.
(1)
To the gray-scale transformation curve Cm of each image-region Pm selection, according to the average brightness value generation of wide area image-region Em.Therefore, even image-region Pm's is big or small less, also can obtain the sampling of enough brightness values.In addition, consequently,, also can select suitable gray-scale transformation curve Cm even for little image-region Pm.
(2)
Select signal correction portion 34,, export the selection signal SS of each pixel by the correction of implementing according to selection signal Sm with the output of image-region unit.Use the specified gray-scale transformation curve candidate G1～Gp of selection signal SS of each pixel, the pixel of the original image that constitutes input signal IS is carried out gray scale handle.Therefore, can obtain to be implemented the output signal OS that more appropriate gray scale is handled.For example can suppress the generation of false contouring.In addition, among the output signal OS, can prevent that the seam on the border of each image-region Pm from becoming obvious artificially.
(3)
Gray scale processing execution portion 25 has the two-dimentional LUT that generates in advance.Therefore, can cut down gray scale and handle needed processing load, in particular, can cut down the needed processing load of generation of gray-scale transformation curve Cm.Consequently, can allow gray scale handle high speed.
(4)
Gray scale processing execution portion 25 uses two-dimentional LUT to carry out gray scale and handles.Here, the content of two-dimentional LUT is read from storage devices such as hard disk that visual processing apparatus 21 had or ROM and is used for gray scale and handled.By changing the content of the two-dimentional LUT that is read, do not need to change the formation of hardware, just can realize various gray scales processing.Also promptly, can realize suitable gray scale processing according to the characteristic of original image.
＜variation 〉
The present invention is not limited in above-mentioned execution mode, can also carry out various distortion in the scope that does not break away from its main points.For example, can with above-mentioned [the 2nd execution mode]＜variation much at one distortion is applied to the 3rd execution mode.[the 2nd execution mode]＜variation particularly〉(10)～(12) in, select signal SS by selecting signal Sm to replace with, CS replaces with output signal OS with the gray scale processing signals, just can similarly use.
Below the distinctive variation of the 3rd execution mode is described.
(1)
In the above-mentioned execution mode, the two-dimentional LUT41 that will be made of the matrix of 64 row, 64 row is as one of two-dimentional LUT example.Here, effect of the present invention is not limited in the two-dimentional LUT of this size.For example, can also be the matrix that is arranged with more gray-scale transformation curve candidates on the line direction.In addition, can also on the matrix column direction, be arranged with the pixel value of the pixel value of input signal IS being divided into the pairing output signal OS of value that obtains after the thinner stride.Specifically, for example can be arranged with the pixel value of output signal OS to each pixel value by the input signal IS of 10 bit representations.
If the size of two-dimentional LUT is bigger, just can carry out more suitable gray scale and handle, if less, just can cut down the memory of preserving two-dimentional LUT.
(2)
What illustrate in the above-mentioned execution mode is, in the value [SS] of selecting signal SS, the footnote that is not gray-scale transformation curve candidate G1～Gp of being had with two-dimentional LUT41 (with reference to Fig. 8) is (under the situation of 1～p) value that equates, in the gray scale of input signal IS is handled, use with the integer of closest value [SS] gray-scale transformation curve candidate G1～Gp as footnote.Here, in the value [SS] of selecting signal SS, the footnote that is not gray-scale transformation curve candidate G1～Gp of being had with two-dimentional LUT41 is (under the situation of 1～p) value that equates, can use with the maximum integer (k) that is no more than the value [SS] of selecting signal SS and be the gray-scale transformation curve candidate Gk of footnote (1≤k≤p-1), with be these both sides of gray-scale transformation curve candidate Gk+1 of footnote with the smallest positive integral (k+1) that surpasses [SS], input signal IS is implemented gray scale to be handled, and the pixel value of the input signal IS after gray scale handled, the later value of decimal point with the value [SS] of selecting signal SS is weighted on average (interior branch), and output signal OS is exported.
(3)
What illustrate in the above-mentioned execution mode is, on the matrix column direction, for example is arranged with the pixel value by high 6 the pairing output signal OS of value of the pixel value of the input signal IS of 10 bit representations.Here, output signal OS also can pass through gray scale processing execution portion 25, as the composition output of the matrix behind low 4 the value enforcement linear interpolation of the pixel value of using input signal IS.Also be, on the matrix column direction, for example be arranged with composition by high 6 the pairing matrix of value of the pixel value of the input signal IS of 10 bit representations, use low 4 value of the pixel value of input signal IS, to the composition of high 6 the pairing matrix of value of the pixel value of input signal IS with give composition that high 6 value of the pixel value of input signal IS adds the pairing matrix of value that obtains after " 1 " (for example, among Fig. 8, composition under 1 row) carries out linear interpolation, and export as output signal OS.
By like this,, also can carry out more suitable gray scale and handle even two-dimentional LUT41's (with reference to Fig. 8) is big or small less.
(4)
What illustrate in the above-mentioned execution mode is, according to the average brightness value of wide area image-region EM, and the pairing selection signal of output image zone Pm Sm.Here, select the output intent of signal Sm to be not limited in this method.For example, can also be according to maximum brightness value or the minimum brightness value of wide area image-region Em, the pairing selection signal of output image zone Pm Sm.In addition, selecting the value [Sm] of signal Sm, can be the average brightness value of wide area image-region Em, maximum brightness value or minimum brightness value itself.
In addition, for example can the pairing selection signal of following output image zone Pm Sm.Also promptly, Pm obtains average brightness value to each image-region, according to each average brightness value, obtains the pairing interim selection signal Sm ' of each image-region Pm.Here, select the value of signal Sm ' temporarily, be the numbering of the footnote of gray-scale transformation curve candidate G1～Gp.And then each image-region that wide area image-region Em is comprised is selected the value of signal Sm ' during average, and as the selection signal Sm of image-region Pm.
(5)
What illustrate in the above-mentioned execution mode is, according to the average brightness value of wide area image-region Em, and the pairing selection signal of output image zone Pm Sm.Here, can be not yet simple average according to wide area image-region Em, but according to weighted average (having the average of weight), the pairing selection signal of output image zone Pm Sm.Use illustrated identical of Figure 11 in detailed content and above-mentioned [the 2nd execution mode], obtain the average brightness value of each image-region that constitutes wide area image-region Em, image-region Ps1, the Ps2 of the average brightness value that the average brightness value that has with image-region Pm is differed greatly ... alleviate its weight, obtain the average brightness value of wide area image-region Em.
By like this, even comprise at wide area image-region Em under the situation (for example wide area image-region Em comprises the situation on the border of the different object of two brightness values) of lightness special region, for the output of selecting signal Sm, the influence that brightness value brought of this special area is also less, thereby can carry out the output of more suitable selection signal Sm.
(6)
In the visual processing apparatus 21, can also have the data of description generating unit, generate data of description as the value that two-dimentional LUT preserved.Specifically, the data of description generating unit is made of with gray-scale transformation curve leading-out portion 10 the image segmentation portion in the visual processing apparatus 1 (with reference to Fig. 1) 2, and the set with a plurality of gray-scale transformation curves of being generated is kept among the two-dimentional LUT as data of description.
In addition, each gray-scale transformation curve of being preserved among the two-dimentional LUT can be associated with the input signal IS that implemented spatial manipulation.In this case, in the visual processing apparatus 21, can and select signal correction portion 24, replace with the spatial manipulation portion that input signal IS is carried out spatial manipulation image segmentation portion 22, selection signal leading-out portion 23.
[the 4th execution mode]
Contrast Figure 22～Figure 25 describes the visual processing apparatus 61 as the 4th execution mode of the present invention.
Visual processing apparatus 61 shown in Figure 22 is devices of visual processes such as the spatial manipulation of carrying out picture signal, gray scale processing.Visual processing apparatus 61 is for example handled in the machine of image, with the device composing images processing unit of the color treatments of carrying out picture signal at computer, television set, digital camera, mobile phone, PDA, printer, scanner etc.
Visual processing apparatus 61 is to use picture signal and the blurred signal that picture signal has been implemented spatial manipulation (blur filter processing) is implemented the device of visual processes, and its feature is arranged in the spatial manipulation.
In the past, when the neighboring pixel that uses object pixel was derived blurred signal, if neighboring pixel comprises the pixel very different with object pixel concentration, blurred signal will be subjected to the influence of the different pixel of concentration.Promptly, in image near the pixel the object edge being carried out under the situation of spatial manipulation, originally is not the influence that the pixel at edge can be subjected to the concentration at edge yet.Therefore, this spatial manipulation for example can cause the problems such as generation of false contouring.
Therefore, require spatial manipulation to adapt with the content of image.Relative therewith, for example the spy opens in the flat 10-75395 communique, generates the different a plurality of blurred signals of fog-level, by synthesizing or switching each blurred signal, exports suitable blurred signal.Be the filter size of change spatial manipulation, the influence of the pixel that inhibition concentration is different by such purpose.
And on the other hand, owing in the above-mentioned communique, generate a plurality of blurred signals, and synthetic or switch each blurred signal, the therefore circuit scale in the device or handle load and increase.
Therefore, as the visual processing apparatus 61 of the 4th execution mode of the present invention, its purpose is: export suitable blurred signal, and circuit scale in the cutting device or processing load.
＜visual processing apparatus 61 〉
What represent among Figure 22 is, picture signal (input signal IS) is implemented visual processes and exported the basic comprising of the visual processing apparatus 61 of visual processes image (output signal OS).Visual processing apparatus 61 has: the brightness value to each pixel of the original image obtained as input signal IS carries out spatial manipulation, and exports the spatial manipulation portion 62 of non-clear signal (unshape) US; And, use pairing input signal IS of same pixel and non-clear signal US, carry out the visual processes of original image, and with the visual processes portion 63 of output signal OS output.
＜spatial manipulation portion 62 〉
Use Figure 23, the spatial manipulation of spatial manipulation portion 62 is described.Spatial manipulation portion 62 from input signal IS, obtains the pixel value of the pixel (below, be called neighboring pixel 66) of the neighboring area of the object pixel 65 of the object that becomes spatial manipulation and object pixel 65.
Neighboring pixel 66 is the pixels that are positioned at the neighboring area of object pixel 65, is to be the pixel that comprises in the neighboring area of horizontal 9 pixels of vertical 9 pixels of center expansion with object pixel 65.In addition, the size of neighboring area is not limited in this situation, can also be littler or bigger.In addition, neighboring pixel 66 according to the distance of distance object pixel 65, is divided into the 1st neighboring pixel 67 and the 2nd neighboring pixel 68 from nearer person.Among Figure 23, the 1st neighboring pixel 67 is to be the pixel that comprises in the zone of horizontal 5 pixels of vertical 5 pixels at center with object pixel 65.In addition, the 2nd neighboring pixel 68 is the pixels that are positioned at the periphery of the 1st neighboring pixel 67.
Spatial manipulation portion 62 carries out filtration operation to object pixel 65.In the filtration operation, use, the pixel value of object pixel 65 with neighboring pixel 66 is weighted on average based on object pixel 65 and the difference of the pixel value of neighboring pixel 66 and the weight of distance.Weighted average is by following formula, and promptly F=(∑ [Wij] * [Aij])/(∑ [Wij]) calculates.Here, [Wij] is in object pixel 65 and neighboring pixel 66, is positioned at the weight coefficient of the pixel of the capable j row of i, and [Aij] is in object pixel 65 and neighboring pixel 66, is positioned at the pixel value of the pixel of the capable j row of i.In addition, " ∑ " expression is each pixel of object pixel 65 and neighboring pixel 66 to be implemented the calculating that adds up to.
Use Figure 24, [Wij] describes to weight coefficient.Weight coefficient [Wij] is the value of determining according to object pixel 65 and the difference and the distance of the pixel value of neighboring pixel 66.In particular, the absolute value of the difference of pixel value is big more, the weight coefficient that the value of giving is more little.In addition, distance is big more gives more little weight coefficient.
For example, for object pixel 65, weight coefficient [Wij] is value [1].
In the 1st neighboring pixel 67, for the absolute value of the difference of the pixel value of pixel value and object pixel 65 less than given threshold value pixel, weight coefficient [Wij] is value [1].In the 1st neighboring pixel 67, greater than for the pixel of given threshold value, weight coefficient [Wij] is value [1/2] for the absolute value of the difference of pixel value.Also promptly, even be included in pixel in the 1st neighboring pixel 67, also different according to the weight coefficient that the difference of pixel value is given.
In the 2nd neighboring pixel 68, less than for the pixel of given threshold value, weight coefficient [Wij] is value [1/2] for the absolute value of the difference of the pixel value of pixel value and object pixel 65.In the 2nd neighboring pixel 68, greater than for the pixel of given threshold value, weight coefficient [Wij] is value [1/4] for the absolute value of the difference of pixel value.Also promptly, even be included in pixel in the 2nd neighboring pixel 68, also different according to the weight coefficient that the difference of pixel value is given.In addition, in 2nd neighboring pixel 68 of distance greater than the 1st neighboring pixel 67 of object pixel 65, be endowed littler weight coefficient.
Here, so-called given threshold value for the pixel value of the object pixel 65 of the value in the scope of getting [0.0～1.0], is big or small values such as value [20/256～60/256].
By the above weighted average that calculates, be used as non-clear signal US output.
＜visual processes portion 63 〉
In the visual processes portion 63, use the value of pairing input signal IS of same pixel and non-clear signal US, carry out visual processes.Here the visual processes of being carried out is that the contrast of input signal IS is strengthened or processing such as dynamic range compression.During contrast is strengthened, use input signal IS, and the signal after will strengthening imposes on input signal IS, to image enforcement sharpening with the difference of non-clear signal US or than the function enhanced signal of strengthening.In the dynamic range compression, from input signal IS, deduct non-clear signal US.
Processing in the visual processes portion 63, can use with input signal IS and non-clear signal US serves as that input and the two-dimentional LUT of output signal output OS carry out.
＜visual processing method program 〉
Above-mentioned processing can be used as visual processing program, by execution such as computers.Visual processing program is to be used for allowing computer carry out the program of the visual processing method of the following stated.
Visual processing method possesses: the brightness value to each pixel of the original image obtained as input signal IS carries out spatial manipulation, and exports the spatial manipulation step of non-clear signal US; And, use pairing input signal IS of same pixel and non-clear signal US, carry out the visual processes of original image, and the visual processes step of output signal output OS.
In the spatial manipulation step, to each pixel of input signal IS, the weighted average described in the explanation of enforcement spatial manipulation portion 62, and export non-clear signal US.Because detailed content is illustrated in front, therefore omit here.
The visual processes step is heavy, uses pairing input signal IS of same pixel and non-clear signal US, the visual processes described in the explanation of enforcement visual processes portion 63, and output signal output OS.Because detailed content is illustrated in front, therefore omit here.
＜effect 〉
Use Figure 25 (a)～(b), the effect of the visual processes that visual processing apparatus 61 is implemented describes.Figure 25 (a) and Figure 25 (b), the processing that expression filter is in the past implemented.Figure 25 (b) represents the processing that filter of the present invention is implemented.
Expression is that neighboring pixel 66 comprises the state of the different object of concentration 71 among Figure 25 (a).In the spatial manipulation of object pixel 65, use smoothing filter with given filtration coefficient.Therefore, not the object pixel 65 of the part of object 71 originally, be subjected to the influence of the concentration of object 71.
Expression is the state of spatial manipulation of the present invention among Figure 25 (b).In the spatial manipulation of the present invention, neighboring pixel 66 is comprised the part 66a of object 71, do not comprise the 1st neighboring pixel 67 of object 71, the 2nd neighboring pixel 68 that does not comprise object 71 and each of object pixel 65, use different weight coefficients to carry out spatial manipulation respectively.Therefore, can suppress the object pixel 65 that spatial manipulation crosses and be subjected to the influence of the very different pixel of concentration, thereby carry out more suitable spatial manipulation.
In addition, in the visual processing apparatus 61, do not need to open flat 10-75395 communique, generate a plurality of blurred signals as the spy.Therefore, the circuit scale in can cutting device or handle load.
In addition, in the visual processing apparatus 61, can allow the shape of image of the filter size of spatial filter and the reference of filter institute in fact, suitably change according to picture material.So, can carry out and the corresponding spatial manipulation of picture material.
＜variation 〉
(1)
The size of above-mentioned neighboring pixel the 66, the 1st neighboring pixel the 67, the 2nd neighboring pixel etc. only is an example, can also be other sizes.
Above-mentioned weight coefficient is an example, can also be other values.For example, can surpass under the situation of given threshold value, given weight coefficient assignment [0] at the absolute value of the difference of pixel value.By like this, the object pixel 65 that can allow spatial manipulation cross is not subjected to the influence of the very different pixel of concentration.This is being in the application of purpose with the contrast reinforcement, and having can be to the script contrast to a certain extent with regard to the undue effect of strengthening of the contrast in the bigger part.
In addition, weight coefficient, the value that can be used as function as follows provides.
(1-a)
Can be the function of variable by absolute value, provide the value of weight coefficient with the difference of pixel value.Function for example hour increases a weight coefficient (near 1) at the absolute value of the difference of pixel value, reduces weight coefficient (near 0) when the absolute value of the difference of pixel value is big, is the dull function that reduces of absolute value about the difference of pixel value.
(1-b)
Can be the function of variable by distance, provide the value of weight coefficient with distance object pixel 65.Function for example increases weight coefficient (near 1) when the close together of distance object pixel 65, reduce weight coefficient (near 0) when the distance of distance object pixel 65 is far away, is the dull function that reduces of distance about distance object pixel 65.
Among above-mentioned (1-a), (1-b), provide weight coefficient more continuously.Therefore, compare, can provide more suitable weight coefficient, suppress excessive contrast and strengthen, suppress the generation of false contouring etc., better handle thereby can carry out visual effect with the situation of using threshold value.
(2)
Above-mentioned processing to each pixel can be that unit carries out with the piece that comprises a plurality of pixels.Specifically, at first, calculate the average pixel value of peripheral piece of the periphery of the average pixel value of object piece of the object that becomes spatial manipulation and object piece.And then, with each average pixel value, use weight coefficient weighted average same as described above.By like this, the average pixel value of object piece is further implemented spatial manipulation.
In this case, spatial manipulation portion 62 usefulness can be elected signal leading-out portion 13 (with reference to Fig. 6) or select signal leading-out portion 23 (with reference to 19).In this case, with [the 2nd execution mode]＜variation〉(6), or [the 3rd execution mode]＜variation identical described in (5).
Below, use Figure 26～Figure 28 to be explained.
" formation "
Figure 26 is the block diagram of expression with the formation of the visual processing apparatus 961 of the illustrated processing of the block unit enforcement usefulness Figure 22～Figure 25 that comprises a plurality of pixels.
Contrast Figure 27 describes the spatial manipulation of spatial manipulation portion 962.The input signal IS that is divided into the image block that comprises a plurality of pixels has been shown among Figure 27.Here, each image block is divided into the zone of 9 pixels that comprise horizontal 3 pixels of vertical 3 pixels.In addition, this dividing method is an example, is not limited in such dividing method.In addition, in order to give full play to the visual processes effect, preferably with sizable zone as object, generate spatial manipulation signal US2.
Among Figure 27, the 1st neighboring area 967 is to be the zone of indulging 3 of 3 horizontal strokes at center with object images piece 965.In addition, the 2nd neighboring area 968 is the zones that are positioned at the periphery of the 1st neighboring area 967.
In the filtration operation, the value of the characteristic parameter of the peripheral image block of object images piece 965 and neighboring area 966 is weighted on average.Here, average weighted weight is determined based on the distance between object images piece 965 and the peripheral image block and the difference of characteristic ginseng value.
More specifically, weighted average is by following formula, and promptly F=(∑ [Wij] * [Aij])/(∑ [Wij]) calculates.
Here, [Wij] is in object images piece 965 and neighboring area 966, be positioned at the pairing weight coefficient of image block of the capable j of i row, [Aij] is in object images piece 965 and neighboring area 966, is positioned at the value of characteristic parameter of the image block of the capable j row of i.In addition, " ∑ " expression is each image block of object images piece 965 and neighboring area 966 to be implemented the calculating that adds up to.
Use Figure 28, [Wij] describes to weight coefficient.
Weight coefficient [Wij] is the determined value of difference according to distance between the peripheral image block of object images piece 965 and neighboring area 966 and characteristic ginseng value.In particular, the absolute value of the difference of characteristic ginseng value is big more, the weight coefficient that the value of giving is more little.In addition, distance is big more gives more little weight coefficient.
For example, for object images piece 965, weight coefficient [Wij] is value [1].
In the 1st neighboring area 967, less than for the peripheral image block of given threshold value, weight coefficient [Wij] is value [1] for the absolute value of the difference of the characteristic ginseng value of characteristic ginseng value and object images piece 965.In the 1st neighboring area 967, greater than for the peripheral image block of given threshold value, weight coefficient [Wij] is value [1/2] for the absolute value of the difference of characteristic ginseng value.Also promptly, even be included in peripheral image block in the 1st neighboring area 967, also different according to the weight coefficient that the difference of characteristic ginseng value is given.
In the 2nd neighboring area 968, less than for the peripheral image block of given threshold value, weight coefficient [Wij] is value [1/2] for the absolute value of the difference of the characteristic ginseng value of characteristic ginseng value and object images piece 965.In the 2nd neighboring area 968, greater than for the peripheral image block of given threshold value, weight coefficient [Wij] is value [1/4] for the absolute value of the difference of characteristic ginseng value.Also promptly, even be included in peripheral image block in the 2nd neighboring area 968, also different according to the weight coefficient that the difference of characteristic ginseng value is given.In addition, in 2nd neighboring area 968 of distance greater than the 1st neighboring area 967 of object images piece 965, be endowed littler weight coefficient.
Here, so-called given threshold value for the characteristic ginseng value of the object images piece 965 of getting the interior value of [0.0～1.0] scope, is big or small values such as value [20/256～60/256].
By the above weighted average that calculates, be used as spatial manipulation signal US2 output.
In the visual processes portion 963, carry out identical visual processes with visual processes portion 63 (with reference to Figure 22).But be with the difference of visual processes portion 63, substitute non-clear signal US, use the spatial manipulation signal US2 of the object images piece of the object pixel that comprises the object that becomes visual processes.In addition, though the processing in the visual processes portion 963, the available object images block unit that comprises object pixel is handled in the lump, also can switch spatial manipulation signal US2 and handle according to the order of the pixel that obtains from input signal IS.
Above processing is carried out all pixels that comprised among the input signal IS.
" effect "
In the processing of spatial manipulation portion 962, implementing with the image block is the processing of unit.Therefore, can cut down the treating capacity of spatial manipulation portion 962, realize visual processes more at a high speed.In addition, can also reduce hardware size.
" variation "
More than explanation is to handle with foursquare block unit.Here, the shape of piece can be arbitrarily.
In addition, above-mentioned weight coefficient, threshold value etc. also can suitably change.
Here, the value of the part of weight coefficient can be value [0].In this case, identical with the effect that is shaped as arbitrary shape that makes neighboring area 966.
In addition, though above explanation is, in the spatial manipulation portion 962, use the characteristic parameter of object images piece 965 and neighboring area 966 to carry out spatial manipulation, spatial manipulation also can only use the characteristic parameter of neighboring area 966 to carry out.Also promptly, in the average weighted weight of spatial manipulation, can be with weight value of being made as [0] of object images piece 965.
(3)
Processing in the visual processes portion 63 is not limited in above.For example, visual processes portion 63, can also use the value A of input signal IS, non-clear signal US value B, dynamic range compression function F 4, strengthen function F 5, will be by following formula, be the value output of the value C that calculates of C=F4 (A) * F5 (A/B) as output signal OS.Here, dynamic range compression function F 4 is convex monotone increasing functions such as power function, for example is expressed as F4 (x)=x^ γ (0＜γ＜1).Strengthening function F 5 is power functions, for example is expressed as F5 (x)=x^ α (0＜α≤1).
In visual processes portion 63, implement under the situation of this processing,, then can compress the dynamic range of input signal IS, strengthen local contrast simultaneously if use suitable non-clear signal US by 62 outputs of spatial manipulation of the present invention portion.
On the other hand, improper at non-clear signal US, and under the fuzzy very few situation, edge strengthening and the suitably not reinforcement of degree of comparing.In addition, under fuzzy too much situation, though the compression of dynamic range is not suitably carried out in the reinforcement of degree of comparing.
[the 5th execution mode]
As the 5th execution mode of the present invention, to the application examples of visual processing apparatus illustrated in above-mentioned the 1st～the 4th execution mode, visual processing method, visual processing program and use its system to describe.
Visual processing apparatus for example is built-in or be connected computer, television set, digital camera, mobile phone, PDA etc. handles in the machine of image, carries out the device of the gray scale processing of image, realizes by integrated circuits such as LSI.
In more detail, each functional block of above-mentioned execution mode both can be distinguished single chip, again can be with its all or part of single chip.In addition, though be example with LSI,, also be called IC, system LSI, super LSI, ultraLSI etc. sometimes here because of the difference of integrated level.
In addition, the method for integrated circuit is not limited in LSI, can also realize by special circuit or general processor.Can also use can be after LSI makes, the FPGA that programmes (FieldProgrammable Gate Array), or use the connection of the circuit unit that can reconstitute LSI inside or the reconfigurable processor of setting.
In addition,, the technology of the integrated circuit of LSI occurred replacing, can certainly use this technology to carry out the integrated of functional block if along with the progress or the derivative other technologies of semiconductor technology.The possibility that also has Applied Biotechnology.
The processing of each piece of Fig. 1, Fig. 6, Figure 19, Figure 22, Figure 26 is for example undertaken by the central arithmetic unit (CPU) that visual processing apparatus has.In addition, be used for implementing the program of each processing, be kept in the storage devices such as hard disk, ROM, read among ROM or the RAM and carry out.In addition, the two-dimentional LUT of institute's reference is stored in the storage devices such as hard disk, ROM in the gray scale processing execution portion 14,25 of Fig. 6, Figure 19, carries out reference as required.In addition, two-dimentional LUT also can provide by direct connection or through the generator that network is connected the two-dimentional LUT on the visual processing apparatus indirectly.In addition, the one dimension LUT of institute's reference is too in the gray scale processing execution portion 44 of Figure 13.
In addition, visual processing apparatus can be built-in or be connected in the machine of processing moving, carries out the device that the gray scale of the image of each frame (each zone) is handled.
In addition, in each visual processing apparatus, carry out illustrated visual processing method in above-mentioned the 1st～the 4th execution mode.
Visual processing program, be built-in or be connected computer, television set, digital camera, mobile phone, PDA etc. and handle in the device in the machine of images, be kept in the storage devices such as hard disk, ROM, the program that the gray scale of carries out image is handled for example provides through medium such as CD-ROM or through network.
What illustrate in the above-mentioned execution mode is that the brightness value of each pixel is handled.Here, the color space of the present invention and input signal IS is irrelevant.Also be, processing in the above-mentioned execution mode, under the situation of input signal IS by expressions such as YCbCr color space, YUV color space, Lab color space, Luv color space, YIQ color space, XYZ color space, YPbPr color space, RGB color spaces, for the brightness (luminance) of each color space, lightness, can use equally.
In addition, under the situation that input signal IS is represented by the RGB color space, the processing in the above-mentioned execution mode can be carried out independently to each composition of RGB.
[the 6th execution mode]
As the 6th execution mode of the present invention, contrast Figure 29～Figure 32 is to the application examples of above-mentioned illustrated visual processing apparatus, visual processing method, visual processing program, and use its system to describe.
Figure 29 provides the block diagram of all formations of the ex100 of system for expression realizes the content of content converting service.The Region Segmentation that provides of communication service is become desirable size, in each unit (cell), be provided as the base station ex107～ex110 of stationary wireless stations respectively.
This content provides the ex100 of system, for example on the Internet ex101, through the ex102 of ISP and telephone network ex104 and base station ex107～ex110, be connected with each machines such as mobile phone ex115 of computer ex111, PDA (personal digital assistant) ex112, video camera ex113, mobile phone ex114, band camera.
But the combination that content provides the ex100 of system to be not limited in Figure 29 can combination in any connect.In addition, each machine also can be without the base station ex107～ex110 as stationary wireless stations, but is directly connected in telephone network ex104.
Video camera ex113, but be the machine of taking moving images such as Digital Video.In addition, mobile phone can be the mobile phone of PDC (Personal Digital Communications) mode, CDMA (CodeDivision Multiple Access) mode, W-CDMA (Wideband-Code DivisionMultiple Access) mode or GSM (Global System for Mobile Communications) mode, or PHS (Personal Handyphone System) etc.
In addition, streaming server ex103 connects video camera ex113 by base station ex109, telephone network ex104, can implement live dispensing etc. based on the data that the encoding process that the user uses video camera ex113 to send is crossed.The encoding process of captured data both can be undertaken by video camera ex113, and also the server that can be handled by the transmission of carrying out data etc. carries out.In addition, the motion image data that camera ex116 is captured, machine ex111 sends to streaming server ex103 as calculated.Camera ex116 is the machine that digital camera etc. can be taken rest image and moving image.In this case, the coding of exercise data both can be undertaken by camera ex116, also can be undertaken by computer ex111.In addition, encoding process is handled in the LSIex117 that computer ex111 or camera ex116 are had.In addition, the software that image encoding/decoding can be used is stored in any medium (CD-ROM, floppy disk, hard disk etc.) as readable medium such as computer ex111.In addition, also can send motion image data by the mobile phone ex115 of band camera.The motion image data of this moment is the data of being crossed by the LSI encoding process that mobile phone ex115 is had.
This content provides among the ex100 of system, the user is to after carrying out encoding process by the content of shootings such as video camera ex113, camera ex116 (for example taking image that music live telecast obtains etc.), send to streaming server ex103, on the other hand, streaming server ex103 banishs to serve for the client of the request of making and states content-data.Client is computer ex111, PDAex112, video camera ex113, the mobile phone ex114 etc. that can decode to the data that encoding process is crossed.By like this, content provider system ex100, data and regeneration that can received code is crossed in client, and then, by in client, receiving and separate code regeneration in real time, become the system that can also realize personal broadcaster.
When carrying out the demonstration of content, can use visual processing apparatus illustrated in the above-mentioned execution mode, visual processing method, visual processing program.For example, computer ex111, PDAex112, video camera ex113, mobile phone ex114 etc. can have the visual processing apparatus shown in the above-mentioned execution mode, and realize visual processing method, visual processing program.
In addition, streaming server ex103 can provide a description data to visual processing apparatus through the Internet ex101.In addition, streaming server ex103 can have many, and different data of description is provided respectively.In addition, streaming server ex103, the generation that can also be described.Like this, can obtain under the situation of data of description through the Internet ex101 at visual processing apparatus, visual processing apparatus does not need to preserve in advance the data of description that is used for visual processes, thus the memory capacity that can cut down visual processing apparatus.In addition, owing to can obtain data of description, therefore can realize different visual processes from a plurality of servers that connect through the Internet ex101.
As an example mobile phone is described.
Figure 30 is the figure that the mobile phone ex115 of the visual processing apparatus with above-mentioned execution mode is described.Mobile phone ex115 has: be used for and base station ex110 between send to receive the antenna ex201 of electric wave; CCD cameras etc. can filmed image or the camera section ex203 of rest image; The display part ex202 of the LCD of the data that demonstration is obtained by the received image of antenna ex201 etc. by the captured image of camera section ex203, decoding etc.; The main part that is constituted by operation keys ex204 group; Be used to carry out the audio output unit ex208 such as loud speaker of voice output; Be used to carry out the sound input part ex205 such as pick-up of sound input; Be used for preserving the data behind the codings such as data of data, moving image or rest image of the data of captured moving image or rest image, the mail that receives or the medium ex207 of decoded data; And, be used for making the socket ex206 that medium ex207 can be installed among the mobile phone ex115.Medium ex207 is that SD card etc. places in plastic housing as can electricity rewriteeing or the nonvolatile memory of deletion be a kind of flash element of EEPROM (Electrically Erasable and Programmable Read Only Memory).
And then contrast Figure 31 describes mobile phone ex115.Among the mobile phone ex115, relatively each one of main part with display part ex202 and operation keys ex204 is carried out the master control part ex311 of Comprehensive Control, power circuit part ex310, operation input control part ex304, the ex312 of image encoding portion, the ex303 of camera interface portion, LCD (Liquid Crystal Display) control part ex302, the ex309 of picture decoding portion, multiplexed separated part ex308, the ex307 of record regenerating portion, department of modulation and demodulation ex306 and the ex305 of acoustic processing portion, ex313 is connected to each other through synchronous bus.
Power circuit part ex310 is after the user makes phone become open state by operation on-hook and power button, by give each power supply from battery pack, with the state of digital mobile phone ex115 starting for working of band camera.
Mobile phone ex115, control based on the master control part ex311 that is constituted with CPU, ROM and RAM etc., when the sound call mode, by the ex305 of acoustic processing portion, to be transformed into digital audio data by the voice signal that sound input part ex205 picks up, and by the ex306 of modulation-demodulation circuit portion it is carried out frequency spectrum diffusion and handle, after the ex301 of transmitter/receiver circuit portion enforcement digitaltoanalogconversion processing and frequency conversion process, send through antenna ex201.In addition, mobile phone ex115, when the sound call mode, to amplifying by the received received signal of antenna ex201, implementing frequency conversion process and analog-to-digital conversion handles, carry out the frequency spectrum back-diffusion by the ex306 of modulation-demodulation circuit portion and handle, be transformed into after the analoging sound signal by the ex305 of acoustic processing portion, ex208 exports it through audio output unit.
In addition, when data communication mode under the situation of send Email, the text data of the Email of the operation push-button ex204 input by operating main body portion, ex304 sends to master control part ex311 through operation input control part.Master control part ex311 carries out frequency spectrum diffusion by department of modulation and demodulation ex306 to text data and handles, and after the ex301 of transmitter/receiver circuit portion enforcement digitaltoanalogconversion processing and frequency conversion process, sends to base station ex110 through antenna ex201.
When data communication mode, send under the situation of view data,, send to the ex312 of image encoding portion through the ex303 of camera interface portion by the captured view data of camera section ex203.In addition, under the situation that does not send view data,, also can directly be presented on the display part ex202 through ex303 of camera interface portion and LCD control part ex302 by the captured view data of camera section ex203.
The ex312 of image encoding portion carries out compressed encoding by the view data that camera section ex203 is supplied with, and is transformed into coded image data, sends it to multiplexed separated part ex308.In addition, meanwhile, mobile phone ex115, the sound that will be picked up by sound input part ex205 in the time of will taking with camera section ex203 sends to multiplexed separated part ex308 through the ex305 of acoustic processing portion as the voice data of numeral.
Multiplexed separated part ex308, the coded image data that the ex312 of image encoding portion is supplied with, the voice data of being supplied with the ex305 of acoustic processing portion, multiplexed in given mode, with the resulting multiplexed data of result, carry out frequency spectrum diffusion by department of modulation and demodulation ex306 and handle, after the ex301 of transmitter/receiver circuit portion enforcement digitaltoanalogconversion processing and frequency conversion process, send through antenna ex201.
When data communication mode, receive under the data conditions of the motion pictures files that is linked on homepage etc., the received signal that will receive from base station ex110 through antenna ex201, implement the frequency spectrum back-diffusion with the ex306 of modulation-demodulation circuit portion and handle, the resulting multiplexed data of result is sent to multiplexed separated part ex308.
In addition, for to decoding through the received multiplexed data of antenna ex201, multiplexed separated part ex308, by separating multiplexed data, tell the coding stream of view data and the coding stream of voice data, through synchronous bus ex313 this coded image data is offered the ex309 of picture decoding portion, simultaneously this voice data is offered the ex305 of acoustic processing portion.
Next, the ex309 of picture decoding portion is by the coding stream of decode image data, generate the regeneration motion image data, it is offered display part ex202 through LCD control part ex302, by like this, demonstrate the motion image data that is comprised in the motion pictures files that is linked on the homepage for example.Meanwhile, the ex305 of acoustic processing portion by voice data being transformed into after the analoging sound signal, offers audio output unit ex208, by like this, bears the voice data that comprises in the motion pictures files that is linked on the homepage for example again.
In the above formation, the ex309 of picture decoding portion can have the visual processing apparatus of above-mentioned execution mode.
In addition, be not limited in the example of said system, recently more and more noticeable based on the digital broadcasting of satellite and surface wave, as shown in Figure 32, digital broadcasting is with also assembling visual processing apparatus illustrated in the above-mentioned execution mode, visual processing method, visual processing program in the system.Specifically, among the ex409 of broadcasting station, the coding stream of image information is sent to communication or broadcasting satellite ex410 through electric wave.Receive its broadcasting satellite ex410, send the broadcasting electric wave, receive this electric wave by the antenna ex406 of family,, coding stream is decoded and regenerated by devices such as television set (receiver) ex401 or set-top box (STB) ex407 with satellite broadcasting receiving equipment.Here, devices such as television set (receiver) ex401 or set-top box (STB) ex407 can have visual processing apparatus illustrated in the above-mentioned execution mode.In addition, also can use the visual processing method of above-mentioned execution mode.In addition, can also have visual processing program.In addition, reading and being stored in medium is coding stream among the medium ex402 such as CD or DVD, and among the regenerating unit ex403 that decodes, the illustrated visual processing apparatus of above-mentioned execution mode, visual processing method, visual processing program can be installed also.In this case, the signal of video signal of being regenerated is presented among the monitor ex404.It is also conceivable that in addition, in connecting the set-top box ex407 of cable TV with the antenna ex406 of cable ex405 or the broadcasting of satellite/terrestrial ripple, visual processing apparatus illustrated in the above-mentioned execution mode, visual processing method, visual processing program are installed, and the monitor ex408 by television set is to its structure of regenerating.At this moment, except in set-top box, can also in television set, assemble visual processing apparatus illustrated in the above-mentioned execution mode.In addition, can also be by having the vehicle ex412 of antenna ex411, from received signals such as satellite ex410 or base station ex107, in the display unit of vehicle mounted guidance ex413 that vehicle ex412 had etc., the regeneration moving image.
In addition, can also coding image signal and be kept in the medium.Object lesson has, the DVD recorder of recording image signal in DVD dish ex421, and the video tape recorder ex420 of the DVR that writes down in hard disk etc.In addition, can also in SD card ex422, carry out record.If video tape recorder ex420 has the decoding device of above-mentioned execution mode, then can coil the picture signal that is write down among ex421 and the SD card ex422 and carry out interpolation regeneration, and be presented among the monitor ex408 DVD.
In addition, formation as vehicle mounted guidance ex413, for example be in the formation shown in Figure 31, remove the structure behind camera section ex203 and the ex303 of camera interface portion, the ex312 of image encoding portion, equally also can consider this structure computer ex111 and television set (receiver) ex401 etc.
In addition, terminals such as above-mentioned mobile phone ex114 except the transmission reception type terminal with these both sides of coder-decoder, it is also conceivable that the transmission terminal that only has encoder, and this three kinds the installation form of receiving terminal that only has decoder.
Like this, illustrated visual processing apparatus, visual processing method, visual processing program in the above-mentioned execution mode can be used in any one above-mentioned machine system, can access effect illustrated in the above-mentioned execution mode.
[note]
The present invention described in the above-mentioned execution mode can also following performance.
The content of＜note 〉
(note 1)
A kind of visual processing apparatus has:
Image region segmentation mechanism is divided into a plurality of image-regions with the picture signal of being imported;
Greyscale transformation characteristic export agency, it is the mechanism that each above-mentioned image-region is derived the greyscale transformation characteristic, its use becomes the gamma characteristic of the ambient image regions of the gamma characteristic of object image area of derived object of above-mentioned greyscale transformation characteristic and above-mentioned object image area, derives the above-mentioned greyscale transformation characteristic of above-mentioned object image area; And,
The gray scale processing mechanism, it is implemented gray scale to above-mentioned picture signal and handles according to the above-mentioned greyscale transformation characteristic that is derived.
(note 2)
According to note 1 described visual processing apparatus:
Above-mentioned greyscale transformation characteristic is a gray-scale transformation curve,
Above-mentioned greyscale transformation characteristic export agency has: use above-mentioned gamma characteristic to generate histogrammic histogram and generate mechanism; And according to the above-mentioned histogram that is generated, the gray-scale transformation curve that generates above-mentioned gray-scale transformation curve generates mechanism.
(note 3)
According to note 1 described visual processing apparatus:
Above-mentioned greyscale transformation characteristic is the selection signal that is used for selecting 1 greyscale transformation table from a plurality of greyscale transformation tables that above-mentioned picture signal carried out the gray scale processing,
Above-mentioned gray scale processing mechanism has above-mentioned a plurality of greyscale transformation table as two-dimentional LUT.
(note 4)
According to note 3 described visual processing apparatus:
Above-mentioned two-dimentional LUT, in whole value of above-mentioned picture signal, dull the increasing or the dull order that reduces of value according to the pairing above-mentioned picture signal of being handled by gray scale of the value of above-mentioned selection signal preserves above-mentioned a plurality of greyscale transformation table.
(note 5)
According to note 3 or 4 described visual processing apparatus:
Above-mentioned two-dimentional LUT can change by the login of data of description.
(note 6)
According to each the described visual processing apparatus in the note 3～5:
The value of above-mentioned selection signal, the characteristic quantity that is used as indivedual selection signals is derived, and this selects signal individually, is the selection signal that each image-region of above-mentioned object image area and above-mentioned ambient image regions is derived.
(note 7)
According to each the described visual processing apparatus in the note 3～5:
Above-mentioned selection signal is derived according to the gamma characteristic characteristic quantity, the characteristic quantity that this gamma characteristic characteristic quantity is to use the gamma characteristic of above-mentioned object image area and above-mentioned ambient image regions to derive.
(note 8)
According to each the described visual processing apparatus in the note 3～7:
Above-mentioned gray scale processing mechanism has: use the selected above-mentioned greyscale transformation table of above-mentioned selection signal, carry out the gray scale processing execution mechanism of the gray scale processing of above-mentioned object image area; And, the mechanism that the gray scale of the above-mentioned picture signal that above-mentioned gray scale was handled is revised, promptly, revise the correction mechanism of the gray scale of above-mentioned object pixel according to comprising the image-region that becomes the object pixel of revising object and comprising the above-mentioned greyscale transformation table that the adjacent image regions of the above-mentioned image-region of above-mentioned object pixel is selected.
(note 9)
According to each the described visual processing apparatus in the note 3～7:
Above-mentioned gray scale processing mechanism has the above-mentioned selection signal of correction, and derivation is used for to the correction mechanism of the correction selection signal of each pixel selection greyscale transformation table of above-mentioned picture signal; And use above-mentioned correction to select the selected above-mentioned greyscale transformation table of signal, carry out the gray scale processing execution mechanism of the gray scale processing of above-mentioned picture signal.
(note 10)
A kind of visual processing method has:
The image region segmentation step is divided into a plurality of image-regions with the picture signal of being imported;
The greyscale transformation characteristic derives step, it is the step that each above-mentioned image-region is derived the greyscale transformation characteristic, its use becomes the gamma characteristic of the ambient image regions of the gamma characteristic of object image area of derived object of above-mentioned greyscale transformation characteristic and above-mentioned object image area, derives the above-mentioned greyscale transformation characteristic of above-mentioned object image area; And,
The gray scale treatment step, it is implemented gray scale to above-mentioned picture signal and handles according to the above-mentioned greyscale transformation characteristic that is derived.
(note 11)
According to note 10 described visual processing methods:
Above-mentioned greyscale transformation characteristic is a gray-scale transformation curve,
Above-mentioned greyscale transformation characteristic derives step, has: use above-mentioned gamma characteristic to generate histogrammic histogram and generate step; And according to the above-mentioned histogram that is generated, the gray-scale transformation curve that generates above-mentioned gray-scale transformation curve generates step.
(note 12)
According to note 10 described visual processing methods:
Above-mentioned greyscale transformation characteristic is the selection signal that is used for selecting 1 greyscale transformation table from a plurality of greyscale transformation tables that above-mentioned picture signal carried out the gray scale processing,
Above-mentioned gray scale treatment step has: use the selected above-mentioned greyscale transformation table of above-mentioned selection signal, carry out the gray scale processing execution step of the gray scale processing of above-mentioned object image area; And, the step that the gray scale of the above-mentioned picture signal that above-mentioned gray scale was handled is revised, promptly, revise the correction step of the gray scale of above-mentioned object pixel according to comprising the image-region that becomes the object pixel of revising object and comprising the above-mentioned greyscale transformation table that the adjacent image regions of the above-mentioned image-region of above-mentioned object pixel is selected.
(note 13)
According to note 10 described visual processing methods:
Above-mentioned greyscale transformation characteristic is the selection signal that is used for selecting 1 greyscale transformation table from a plurality of greyscale transformation tables that above-mentioned picture signal carried out the gray scale processing;
Above-mentioned gray scale treatment step has the above-mentioned selection signal of correction, and derivation is used for to the correction step of the correction selection signal of each pixel selection greyscale transformation table of above-mentioned picture signal; And use above-mentioned correction to select the selected above-mentioned greyscale transformation table of signal, carry out the gray scale processing execution step of the gray scale processing of above-mentioned picture signal.
(note 14)
A kind of visual processing program is used for by computer-implemented visual processing method,
The visual processing method that above-mentioned visual processing program is carried out in computer comprises:
The image region segmentation step is divided into a plurality of image-regions with the picture signal of being imported;
The greyscale transformation characteristic derives step, it is the step that each above-mentioned image-region is derived the greyscale transformation characteristic, its use becomes the gamma characteristic of the ambient image regions of the gamma characteristic of object image area of derived object of above-mentioned greyscale transformation characteristic and above-mentioned object image area, derives the above-mentioned greyscale transformation characteristic of above-mentioned object image area; And,
The gray scale treatment step, it is implemented gray scale to above-mentioned picture signal and handles according to the above-mentioned greyscale transformation characteristic that is derived.
(note 15)
According to note 14 described visual processing programs:
Above-mentioned greyscale transformation characteristic is a gray-scale transformation curve,
Above-mentioned greyscale transformation characteristic derives step, has: use above-mentioned gamma characteristic to generate histogrammic histogram and generate step; And according to the above-mentioned histogram that is generated, the gray-scale transformation curve that generates above-mentioned gray-scale transformation curve generates step.
(note 16)
According to note 14 described visual processing programs:
Above-mentioned greyscale transformation characteristic is the selection signal that is used for selecting 1 greyscale transformation table from a plurality of greyscale transformation tables that above-mentioned picture signal carried out the gray scale processing,
Above-mentioned gray scale treatment step has: use the selected above-mentioned greyscale transformation table of above-mentioned selection signal, carry out the gray scale processing execution step of the gray scale processing of above-mentioned object image area; And, the step that the gray scale of the above-mentioned picture signal that above-mentioned gray scale was handled is revised, promptly, revise the correction step of the gray scale of above-mentioned object pixel according to comprising the image-region that becomes the object pixel of revising object and comprising the above-mentioned greyscale transformation table that the adjacent image regions of the above-mentioned image-region of above-mentioned object pixel is selected.
(note 17)
According to note 14 described visual processing programs:
Above-mentioned greyscale transformation characteristic is the selection signal that is used for selecting 1 greyscale transformation table from a plurality of greyscale transformation tables that above-mentioned picture signal carried out the gray scale processing;
Above-mentioned gray scale treatment step has the above-mentioned selection signal of correction, and derivation is used for to the correction step of the correction selection signal of each pixel selection greyscale transformation table of above-mentioned picture signal; And use above-mentioned correction to select the selected above-mentioned greyscale transformation table of signal, carry out the gray scale processing execution step of the gray scale processing of above-mentioned picture signal.
(note 18)
According to note 1 described visual processing apparatus:
Above-mentioned gray scale processing mechanism, have and to be used for above-mentioned picture signal is carried out the parameter of curve of the gray-scale transformation curve that gray scale handles, parameter output mechanism according to above-mentioned greyscale transformation characteristic output, use is carried out gray scale to above-mentioned picture signal and is handled according to above-mentioned greyscale transformation characteristic and the determined above-mentioned gray-scale transformation curve of above-mentioned parameter of curve.
(note 19)
According to note 18 described visual processing apparatus:
The above-mentioned parameter output mechanism is a look-up table of preserving the relation between above-mentioned greyscale transformation characteristic and the above-mentioned parameter of curve.
(note 20)
According to note 18 or 19 described visual processing apparatus:
Above-mentioned parameter of curve, the set-point that comprises above-mentioned picture signal is pairing by the value of above-mentioned gray scale processed images signal.
(note 21)
According to each the described visual processing apparatus in the note 18～20:
Above-mentioned parameter of curve, the slope of the above-mentioned gray-scale transformation curve in comprising between the given area of above-mentioned picture signal.
(note 22)
According to each the described visual processing apparatus in the note 18～21:
Above-mentioned parameter of curve comprises at least 1 the coordinate that above-mentioned gray-scale transformation curve passes through.
(note 23)
A kind of visual processing apparatus has:
Spatial manipulation mechanism, be that each of a plurality of image-regions in the picture signal of being imported is carried out spatial manipulation, and the mechanism of derived space processing signals, in the above-mentioned spatial manipulation, use based on the gamma characteristic of the object image area of the object that becomes above-mentioned spatial manipulation, with the weight of the difference of the gamma characteristic of the ambient image regions of above-mentioned object image area, carry out the weighted average of the gamma characteristic of above-mentioned object image area and above-mentioned ambient image regions; And,
Visual processes mechanism, it carries out the visual processes of above-mentioned object image area according to the gamma characteristic and the above-mentioned spatial manipulation signal of above-mentioned object image area.
(note 24)
According to note 23 described visual processing apparatus:
The absolute value of the difference of above-mentioned gamma characteristic is big more, and above-mentioned weighting is just more little.
(note 25)
According to note 23 or 24 described visual processing apparatus:
Distance between above-mentioned object image area and the above-mentioned ambient image regions is big more, and above-mentioned weighting is just more little.
(note 26)
According to each the described visual processing apparatus in the note 23～25:
Above-mentioned image-region is made of a plurality of pixels;
The gamma characteristic of above-mentioned object image area and above-mentioned ambient image regions, the characteristic quantity that is used as the pixel value that constitutes each image-region determines.
(note 27)
A kind of visual processing apparatus has:
The object image area determination means, it is from the picture signal of being imported, and decision becomes the object image area of the derived object of greyscale transformation characteristic;
The ambient image regions determination means, its decision is positioned at the periphery of above-mentioned object image area, and comprises at least 1 ambient image regions of a plurality of pixels;
Greyscale transformation characteristic export agency, the peripheral view data that it uses above-mentioned ambient image regions derives the above-mentioned greyscale transformation characteristic of above-mentioned object image area; And,
The gray scale processing mechanism, it is implemented gray scale to the picture signal of above-mentioned object image area and handles according to the above-mentioned greyscale transformation characteristic that is derived.
(note 28)
A kind of visual processing method has:
The object image area deciding step, it is from the picture signal of being imported, and decision becomes the object image area of the derived object of greyscale transformation characteristic;
The ambient image regions deciding step, its decision is positioned at the periphery of above-mentioned object image area, and comprises at least 1 ambient image regions of a plurality of pixels;
The greyscale transformation characteristic derives step, and the peripheral view data that it uses above-mentioned ambient image regions derives the above-mentioned greyscale transformation characteristic of above-mentioned object image area; And,
The gray scale treatment step, it is implemented gray scale to the picture signal of above-mentioned object image area and handles according to the above-mentioned greyscale transformation characteristic that is derived.
(note 29)
A kind of visual processing program is used to use a computer and carries out the visual processing method that the picture signal of being imported is carried out visual processes,
Above-mentioned visual processing method comprises:
The object image area deciding step, it is from the picture signal of being imported, and decision becomes the object image area of the derived object of greyscale transformation characteristic;
The ambient image regions deciding step, its decision is positioned at the periphery of above-mentioned object image area, and comprises at least 1 ambient image regions of a plurality of pixels;
The greyscale transformation characteristic derives step, and the peripheral view data that it uses above-mentioned ambient image regions derives the above-mentioned greyscale transformation characteristic of above-mentioned object image area; And,
The gray scale treatment step, it is implemented gray scale to the picture signal of above-mentioned object image area and handles according to the above-mentioned greyscale transformation characteristic that is derived.
(note 30)
A kind of semiconductor device has:
The object image area determination section, it is from the picture signal of being imported, and decision becomes the object image area of the derived object of greyscale transformation characteristic;
The ambient image regions determination section, its decision is positioned at the periphery of above-mentioned object image area, and comprises at least 1 ambient image regions of a plurality of pixels;
Greyscale transformation characteristic leading-out portion, the peripheral view data that it uses above-mentioned ambient image regions derives the above-mentioned greyscale transformation characteristic of above-mentioned object image area; And,
The gray scale handling part, it is implemented gray scale to the picture signal of above-mentioned object image area and handles according to the above-mentioned greyscale transformation characteristic that is derived.
(explanation of note)
Visual processing apparatus described in the note 1 has: image region segmentation mechanism, greyscale transformation characteristic export agency and gray scale processing mechanism.Image region segmentation mechanism is divided into a plurality of image-regions with the picture signal of being imported.Greyscale transformation characteristic export agency, it is the mechanism of each image-region being derived the greyscale transformation characteristic, its use become the greyscale transformation characteristic derived object object image area gamma characteristic, with the gamma characteristic of the ambient image regions of object image area, the greyscale transformation characteristic of derived object image-region.The gray scale processing mechanism according to the greyscale transformation characteristic that is derived, carries out the gray scale of picture signal and handles.
Here, the greyscale transformation characteristic is meant, the characteristic that the gray scale of each image-region is handled.So-called gamma characteristic for example is the pixel values such as brightness, lightness of each pixel.
In the visual processing apparatus of the present invention, when judging the greyscale transformation characteristic of each image-region,, also use the gamma characteristic of the wide area image-region of the image-region that comprises periphery to judge except the gamma characteristic of each image-region.Therefore, can handle the effect of adding spatial manipulation, thereby can realize further improving the gray scale processing of visual effect the gray scale of each image-region.
Visual processing apparatus described in the note 2 is a kind of according to note 1 described visual processing apparatus, and wherein, the greyscale transformation characteristic is a gray-scale transformation curve.Greyscale transformation characteristic export agency has and use gamma characteristic to generate histogrammic histogram to generate mechanism, and according to the histogram that is generated, the gray-scale transformation curve that generates gray-scale transformation curve generates mechanism.
Here, so-called histogram for example is meant the pairing distribution of gamma characteristic of the pixel that object image area and ambient image regions comprise.Gray-scale transformation curve generates mechanism, for example will accumulate the cumulative curve of histogrammic value as gray-scale transformation curve.
In the visual processing apparatus of the present invention, when generating histogram, except the gamma characteristic of each image-region, also use the gamma characteristic of the wide area that comprises ambient image regions, carry out histogrammic generation.Therefore, can increase the number of cutting apart of picture signal, reduce the size of image-region, suppress the generation that gray scale is handled the false contouring that causes.In addition, the border that can prevent image-region becomes obvious artificially.
Visual processing apparatus described in the note 3 is a kind of according to note 1 described visual processing apparatus, and wherein, the greyscale transformation characteristic is the selection signal that is used for selecting 1 greyscale transformation table from a plurality of greyscale transformation tables that picture signal carried out the gray scale processing.The gray scale processing mechanism has a plurality of greyscale transformation tables as two-dimentional LUT.
Here, so-called greyscale transformation table for example is the look-up table (LUT) etc. of the pixel value of the corresponding image signals pixel value of preserving gray scale processed images signal.
Select the value of signal, for example the value of distributing to 1 greyscale transformation table for from each the value of distributing to a plurality of greyscale transformation tables, selecting.The gray scale processing mechanism, according to the value of selecting signal and the pixel value of picture signal, the pixel value of the picture signal of gray scale processing was implemented in output with reference to two-dimentional LUT.
In the visual processing apparatus of the present invention, carry out gray scale with reference to two-dimentional LUT and handle.Therefore, can make gray scale handle high speed.In addition, handle, therefore can carry out suitable gray scale and handle owing to from a plurality of greyscale transformation tables, select 1 greyscale transformation table to carry out gray scale.
Visual processing apparatus described in the note 4, be a kind of according to note 3 described visual processing apparatus, wherein, two dimension LUT, in whole value of picture signal, dull the increasing or the dull order that reduces of value according to the value corresponding gray scale processed images signal of selecting signal preserves a plurality of greyscale transformation tables.
In the visual processing apparatus of the present invention, for example, allow the value of selecting signal demonstrate the degree of greyscale transformation.
Visual processing apparatus described in the note 5 is a kind of according to note 3 or 4 described visual processing apparatus, and wherein, two-dimentional LUT can change by the login of data of description.
Here, data of description is meant the data that are kept among the two-dimentional LUT, and for example the pixel value with gray scale processed images signal is a key element.
In the visual processing apparatus of the present invention,, do not need to change hardware and just constitute and to carry out various changes to the characteristic that gray scale is handled by changing two-dimentional LUT.
Visual processing apparatus described in the note 6, be a kind of according to each the described visual processing apparatus in the note 3～5, wherein, select the value of signal, as the characteristic quantity of the promptly indivedual selection signals of the selection signal that each image-region of object image area and ambient image regions is derived and derive.
Here, selecting the characteristic quantity of signal individually, for example is mean value (simple average or weighted average), maximum or minimum value to the selection signal of each image-region derivation.
In the visual processing apparatus of the present invention,, derive as the characteristic quantity of the pairing selection signal of wide area image-region that comprises ambient image regions with the pairing selection signal of object image area.Therefore, can add the effect of spatial manipulation to the selection signal, thereby the border that can prevent image-region becomes obvious artificially.
Visual processing apparatus described in the note 7, be a kind of according to each the described visual processing apparatus in the note 3～5, wherein, select signal, the characteristic quantity of deriving according to the gamma characteristic of using object image area and ambient image regions is that the gamma characteristic characteristic quantity is derived.
Here, the gamma characteristic characteristic quantity for example is mean value (simple average or weighted average), maximum or the minimum value etc. of gamma characteristic of the wide area of object image area and ambient image regions.
In the visual processing apparatus of the present invention,, derive according to the wide area image-region corresponding gray scale property feature amount that comprises ambient image regions with the pairing selection signal of object image area.Therefore, can add the effect of spatial manipulation to the selection signal, thereby the border that can prevent image-region becomes obvious artificially.
Visual processing apparatus described in the note 8 is a kind of according to each the described visual processing apparatus in the note 3～7, and wherein, the gray scale processing mechanism has gray scale processing execution mechanism, and correction mechanism.Gray scale processing execution mechanism uses and selects the selected greyscale transformation table of signal, carries out the gray scale of object image area and handles.Correction mechanism, it is the mechanism that the gray scale of gray scale processed images signal is revised, according to the greyscale transformation table that the adjacent image regions that comprises image-region that becomes the object pixel of revising object and the image-region that comprises object pixel is selected, revise the gray scale of object pixel.
Here, adjacent image regions both can be the identical image-region of ambient image regions when deriving the greyscale transformation characteristic, also can be different image-regions.For example, adjacent image regions can be chosen 3 the shortest image-regions of distance of object pixel in the image-region adjacent with the image-region that comprises object pixel.
Correction mechanism for example according to each object image area, is implemented to revise to the gray scale of using identical greyscale transformation table gray scale processed images signal.The correction of object pixel is for example implemented according to the position of object pixel, makes the influence of each greyscale transformation table that adjacent image regions is selected show.
In the visual processing apparatus of the present invention, can be to the gray scale of each pixel correction picture signal.Therefore, can prevent further that the border of image-region from becoming eye-catching artificially, thereby can improve visual effect.
Visual processing apparatus described in the note 9 is a kind of according to each the described visual processing apparatus in the note 3～7, and wherein, the gray scale processing mechanism has correction mechanism and gray scale processing execution mechanism.Correction mechanism is revised and is selected signal, each pixel derivation of picture signal is used for selecting the correction selection signal of greyscale transformation table.Gray scale processing execution mechanism uses to revise and selects the selected greyscale transformation table of signal, and the gray scale of carries out image signal is handled.
Correction mechanism for example will be revised according to location of pixels and to the selection signal that the image-region adjacent with object image area derived, and be derived the selection signal of each pixel the selection signal of each object image area derivation.
In the visual processing apparatus of the present invention, can derive each pixel and select signal.Therefore, can prevent further that the border of image-region from becoming eye-catching artificially, thereby can improve visual effect.
Visual processing method described in the note 10 has: image region segmentation step, greyscale transformation characteristic derive step and gray scale treatment step.The image region segmentation step is divided into a plurality of image-regions with the picture signal of being imported.The greyscale transformation characteristic derives step, it is the step that each image-region is derived the greyscale transformation characteristic, its use become the greyscale transformation characteristic derived object object image area gamma characteristic, with the gamma characteristic of the ambient image regions of object image area, the greyscale transformation characteristic of derived object image-region.The gray scale treatment step according to the greyscale transformation characteristic that is derived, carries out the gray scale of picture signal and handles.
Here, the greyscale transformation characteristic is meant, the characteristic that the gray scale of each image-region is handled.So-called gamma characteristic for example is the pixel values such as brightness, lightness of each pixel.
In the visual processing method of the present invention, when judging the greyscale transformation characteristic of each image-region,, also use the gamma characteristic of the wide area image-region of the image-region that comprises periphery to judge except the gamma characteristic of each image-region.Therefore, can handle the effect of adding spatial manipulation, thereby can realize further improving the gray scale processing of visual effect the gray scale of each image-region.
Visual processing method described in the note 11 is a kind of according to note 10 described visual processing methods, and wherein, the greyscale transformation characteristic is a gray-scale transformation curve.The greyscale transformation characteristic derives step, have the gamma characteristic of use and generate histogrammic histogram generation step, and according to the histogram that is generated, the gray-scale transformation curve that generates gray-scale transformation curve generates step.
Here, so-called histogram for example is meant the pairing distribution of gamma characteristic of the pixel that object image area and ambient image regions comprise.Gray-scale transformation curve generates step, for example will accumulate the cumulative curve of histogrammic value as gray-scale transformation curve.
In the visual processing method of the present invention, when generating histogram, except the gamma characteristic of each image-region, also use the gamma characteristic of the wide area that comprises ambient image regions, carry out histogrammic generation.Therefore, can increase the number of cutting apart of picture signal, reduce the size of image-region, suppress the generation that gray scale is handled the false contouring that causes.In addition, the border that can prevent image-region becomes obvious artificially.
Visual processing method described in the note 12 is a kind of according to note 10 described visual processing methods, and wherein, the greyscale transformation characteristic is the selection signal that is used for selecting 1 greyscale transformation table from a plurality of greyscale transformation tables that picture signal carried out the gray scale processing.In addition, the gray scale treatment step has gray scale processing execution step, and revises step.Gray scale processing execution step is used and is selected the selected greyscale transformation table of signal, carries out the gray scale of object image area and handles.Revise step, it is the step that the gray scale of gray scale processed images signal is revised, according to the greyscale transformation table that the adjacent image regions that comprises image-region that becomes the object pixel of revising object and the image-region that comprises object pixel is selected, revise the gray scale of object pixel.
Here, so-called greyscale transformation table for example is the look-up table (LUT) etc. of the pixel value of the corresponding image signals pixel value of preserving gray scale processed images signal.Adjacent image regions both can be the identical image-region of ambient image regions when deriving the greyscale transformation characteristic, also can be different image-regions.For example, adjacent image regions can be chosen 3 the shortest image-regions of distance of object pixel in the image-region adjacent with the image-region that comprises object pixel.
Select the value of signal, for example the value of distributing to 1 greyscale transformation table for from each the value of distributing to a plurality of greyscale transformation tables, selecting.The gray scale treatment step, according to the value of selecting signal and the pixel value of picture signal, the pixel value of the picture signal of gray scale processing was implemented in output with reference to two-dimentional LUT.Revise step,, the gray scale of using identical greyscale transformation table gray scale processed images signal is implemented to revise for example according to each object image area.The correction of object pixel is for example implemented according to the position of object pixel, makes the influence of each greyscale transformation table that adjacent image regions is selected show.
In the visual processing method of the present invention, carry out gray scale with reference to LUT and handle.Therefore, can make gray scale handle high speed.In addition, handle, therefore can carry out suitable gray scale and handle owing to from a plurality of greyscale transformation tables, select 1 greyscale transformation table to carry out gray scale.In addition, can be to the gray scale of each pixel correction picture signal.Therefore, can prevent further that the border of image-region from becoming eye-catching artificially, thereby can improve visual effect.
Visual processing method described in the note 13 is a kind of according to note 10 described visual processing methods, and wherein, the greyscale transformation characteristic is the selection signal that is used for selecting 1 greyscale transformation table from a plurality of greyscale transformation tables that picture signal carried out the gray scale processing.In addition, the gray scale treatment step has the step of correction and gray scale processing execution step.Revise step, revise and select signal, each pixel derivation of picture signal is used for selecting the correction selection signal of greyscale transformation table.Gray scale processing execution step is used to revise and is selected the selected greyscale transformation table of signal, and the gray scale of carries out image signal is handled.
Here, so-called greyscale transformation table for example is the look-up table (LUT) etc. of the pixel value of the corresponding image signals pixel value of preserving gray scale processed images signal.
Select the value of signal, for example the value of distributing to 1 greyscale transformation table for from each the value of distributing to a plurality of greyscale transformation tables, selecting.The gray scale treatment step, according to the value of selecting signal and the pixel value of picture signal, the pixel value of the picture signal of gray scale processing was implemented in output with reference to two-dimentional LUT.Revise step, for example will the selection signal of each object image area derivation be revised according to location of pixels and to the selection signal that the image-region adjacent with object image area derived, and derive the selection signal of each pixel.
Visual processing method of the present invention carries out gray scale with reference to LUT and handles.Therefore, can make gray scale handle high speed.In addition, handle, therefore can carry out suitable gray scale and handle owing to from a plurality of greyscale transformation tables, select 1 greyscale transformation table to carry out gray scale.In addition, can derive the selection signal to each pixel.Therefore, can prevent further that the border of image-region from becoming eye-catching artificially, thereby can improve visual effect.
Visual processing program described in the note 14 is by computer, and execution has the visual processing program of the visual processing method of image region segmentation step, greyscale transformation characteristic derivation step and gray scale treatment step.The image region segmentation step is divided into a plurality of image-regions with the picture signal of being imported.The greyscale transformation characteristic derives step, it is the step that each image-region is derived the greyscale transformation characteristic, its use become the greyscale transformation characteristic derived object object image area gamma characteristic, with the gamma characteristic of the ambient image regions of object image area, the greyscale transformation characteristic of derived object image-region.The gray scale treatment step according to the greyscale transformation characteristic that is derived, carries out the gray scale of picture signal and handles.
Here, the greyscale transformation characteristic is meant, the characteristic that the gray scale of each image-region is handled.So-called gamma characteristic for example is the pixel values such as brightness, lightness of each pixel.
In the visual processing program of the present invention, when judging the greyscale transformation characteristic of each image-region,, also use the gamma characteristic of the wide area image-region of the image-region that comprises periphery to judge except the gamma characteristic of each image-region.Therefore, can handle the effect of adding spatial manipulation, thereby can realize further improving the gray scale processing of visual effect the gray scale of each image-region.
Visual processing program described in the note 15 is a kind of according to note 14 described visual processing programs, and wherein, the greyscale transformation characteristic is a gray-scale transformation curve.The greyscale transformation characteristic derives step, have the gamma characteristic of use and generate histogrammic histogram generation step, and according to the histogram that is generated, the gray-scale transformation curve that generates gray-scale transformation curve generates step.
Here, so-called histogram for example is meant the pairing distribution of gamma characteristic of the pixel that object image area and ambient image regions comprise.Gray-scale transformation curve generates step, for example will accumulate the cumulative curve of histogrammic value as gray-scale transformation curve.
In the visual processing program of the present invention, when generating histogram, except the gamma characteristic of each image-region, also use the gamma characteristic of the wide area that comprises ambient image regions, carry out histogrammic generation.Therefore, can increase the number of cutting apart of picture signal, reduce the size of image-region, suppress the generation that gray scale is handled the false contouring that causes.In addition, the border that can prevent image-region becomes obvious artificially.
Visual processing program described in the note 16 is a kind of according to note 14 described visual processing programs, and wherein, the greyscale transformation characteristic is the selection signal that is used for selecting 1 greyscale transformation table from a plurality of greyscale transformation tables that picture signal carried out the gray scale processing.In addition, the gray scale treatment step has gray scale processing execution step, and revises step.Gray scale processing execution step is used and is selected the selected greyscale transformation table of signal, carries out the gray scale of object image area and handles.Revise step, it is the step that the gray scale of gray scale processed images signal is revised, according to the greyscale transformation table that the adjacent image regions that comprises image-region that becomes the object pixel of revising object and the image-region that comprises object pixel is selected, revise the gray scale of object pixel.
Here, so-called greyscale transformation table for example is the look-up table (LUT) etc. of the pixel value of the corresponding image signals pixel value of preserving gray scale processed images signal.Adjacent image regions both can be the identical image-region of ambient image regions when deriving the greyscale transformation characteristic, also can be different image-regions.For example, adjacent image regions can be chosen 3 the shortest image-regions of distance of object pixel in the image-region adjacent with the image-region that comprises object pixel.
Select the value of signal, for example the value of distributing to 1 greyscale transformation table for from each the value of distributing to a plurality of greyscale transformation tables, selecting.The gray scale treatment step, according to the value of selecting signal and the pixel value of picture signal, the pixel value of the picture signal of gray scale processing was implemented in output with reference to two-dimentional LUT.Revise step,, the gray scale of using identical greyscale transformation table gray scale processed images signal is implemented to revise for example according to each object image area.The correction of object pixel is for example implemented according to the position of object pixel, makes the influence of each greyscale transformation table that adjacent image regions is selected show.
In the visual processing program of the present invention, carry out gray scale with reference to LUT and handle.Therefore, can make gray scale handle high speed.In addition, handle, therefore can carry out suitable gray scale and handle owing to from a plurality of greyscale transformation tables, select 1 greyscale transformation table to carry out gray scale.In addition, can be to the gray scale of each pixel correction picture signal.Therefore, can prevent further that the border of image-region from becoming eye-catching artificially, thereby can improve visual effect.
Visual processing program described in the note 17 is a kind of according to note 14 described visual processing programs, and wherein, the greyscale transformation characteristic is the selection signal that is used for selecting 1 greyscale transformation table from a plurality of greyscale transformation tables that picture signal carried out the gray scale processing.In addition, the gray scale treatment step has the step of correction and gray scale processing execution step.Revise step, revise and select signal, each pixel derivation of picture signal is used for selecting the correction selection signal of greyscale transformation table.Gray scale processing execution step is used to revise and is selected the selected greyscale transformation table of signal, and the gray scale of carries out image signal is handled.
Here, so-called greyscale transformation table for example is the look-up table (LUT) etc. of the pixel value of the corresponding image signals pixel value of preserving gray scale processed images signal.
Select the value of signal, for example the value of distributing to 1 greyscale transformation table for from each the value of distributing to a plurality of greyscale transformation tables, selecting.The gray scale treatment step, according to the value of selecting signal and the pixel value of picture signal, the pixel value of the picture signal of gray scale processing was implemented in output with reference to two-dimentional LUT.Revise step, for example will the selection signal of each object image area derivation be revised according to location of pixels and to the selection signal that the image-region adjacent with object image area derived, and derive the selection signal of each pixel.
Visual processing program of the present invention carries out gray scale with reference to LUT and handles.Therefore, can make gray scale handle high speed.In addition, handle, therefore can carry out suitable gray scale and handle owing to from a plurality of greyscale transformation tables, select 1 greyscale transformation table to carry out gray scale.In addition, can derive the selection signal to each pixel.Therefore, can prevent further that the border of image-region from becoming eye-catching artificially, thereby can improve visual effect.
Visual processing apparatus described in the note 18, be a kind of according to note 1 described visual processing apparatus, wherein, the gray scale processing mechanism, have and to be used for picture signal is carried out the parameter of curve of the gray-scale transformation curve that gray scale handles, according to the parameter output mechanism of greyscale transformation characteristic output.The gray scale processing mechanism uses according to greyscale transformation characteristic and the determined gray-scale transformation curve of parameter of curve, picture signal is carried out gray scale handle.
Here, gray-scale transformation curve, also comprising at least a portion is the curve of straight line.Parameter of curve is to be used for allowing gray-scale transformation curve and other other parameter of gray-scale transformation curve phase region, for example slope of the coordinate on the gray-scale transformation curve, gray-scale transformation curve, curvature etc.The parameter output mechanism for example is a look-up table of preserving the pairing parameter of curve of greyscale transformation characteristic, perhaps by using the pairing parameter of curve of given greyscale transformation characteristic to implement computings such as curve approximation, obtains the arithmetical organ of parameter of curve etc.
In the visual processing apparatus of the present invention, according to the greyscale transformation characteristic picture signal is carried out gray scale and handle.Therefore, can carry out more suitable gray scale handles.In addition, do not need to preserve in advance the value of employed all gray-scale transformation curves in the gray scale processing, determine gray-scale transformation curve, carry out gray scale and handle according to the parameter of curve of being exported.Therefore, can cut down the memory capacity that is used for preserving gray-scale transformation curve.
Visual processing apparatus described in the note 19 is a kind of according to note 18 described visual processing apparatus, and wherein, the parameter output mechanism is a look-up table of preserving the relation between greyscale transformation characteristic and the parameter of curve.
Look-up table is preserved the relation between greyscale transformation characteristic and the parameter of curve.The gray scale processing mechanism uses determined gray-scale transformation curve, picture signal is carried out gray scale handle.
In the visual processing apparatus of the present invention, according to the greyscale transformation characteristic picture signal is carried out gray scale and handle.Therefore, can carry out more suitable gray scale handles.In addition, do not need to preserve the value of employed all gray-scale transformation curves in advance, and only preserve parameter of curve.Therefore, can cut down the memory capacity that is used for preserving gray-scale transformation curve.
Visual processing apparatus described in the note 20 is a kind of according to note 18 or 19 described visual processing apparatus, and wherein, parameter of curve comprises the value of the set-point corresponding gray scale processed images signal of picture signal.
In the gray scale processing mechanism, use the set-point of picture signal and become relation between the value of picture signal of object of visual processes, the value of the gray scale processed images signal that parameter of curve is comprised carries out dividing in non-linear or linear, derives the value of gray scale processed images signal.
In the visual processing apparatus of the present invention, can determine gray-scale transformation curve, carry out gray scale and handle according to the pairing value of having implemented the picture signal of gray scale processing of the set-point of picture signal.
Visual processing apparatus described in the note 21 is a kind of according to each the described visual processing apparatus in the note 18～20, wherein, and parameter of curve, the slope of the gray-scale transformation curve in comprising between the given area of picture signal.
The gray scale processing mechanism by the slope of the gray-scale transformation curve between the given area of picture signal, is determined gray-scale transformation curve.In addition, use determined gray-scale transformation curve, the pairing value of implementing the picture signal of gray scale processing of the value of deduced image signal.
In the visual processing apparatus of the present invention, can determine gray-scale transformation curve, carry out gray scale and handle according to the slope of the gray-scale transformation curve between the given area of picture signal.
Visual processing apparatus described in the note 22 is a kind of according to each the described visual processing apparatus in the note 18～21, and wherein, parameter of curve comprises at least 1 the coordinate that gray-scale transformation curve passes through.
In the parameter of curve, determine at least 1 coordinate by gray-scale transformation curve.At least 1 point of the value of the picture signal after also promptly, the value corresponding gray scale of determining picture signal is handled.The gray scale processing mechanism, use determined picture signal value, and become relation between the value of picture signal of object of visual processes, be divided in the value of the picture signal after determined gray scale handled non-linear or linear, by the value of such derivation gray scale processed images signal.
In the visual processing apparatus of the present invention, can determine gray-scale transformation curve, carry out gray scale and handle according at least 1 the coordinate that gray-scale transformation curve passed through.
Visual processing apparatus described in the note 23 has spatial manipulation mechanism and visual processes mechanism.Spatial manipulation mechanism is that each of a plurality of image-regions in the picture signal of being imported is carried out spatial manipulation, the mechanism of derived space processing signals.In the spatial manipulation, use based on the gamma characteristic of the object image area of the object that becomes spatial manipulation, with the weighting of the difference of the gamma characteristic of the ambient image regions of object image area, carry out the weighted average of the gamma characteristic of object image area and ambient image regions.Visual processes mechanism according to the gamma characteristic and the spatial manipulation signal of object image area, carries out the visual processes of object image area.
Here, image-region is meant, comprises the zone of a plurality of pixels in image, or pixel itself.Gamma characteristic is meant, based on the value of pixel values such as the brightness of each pixel, lightness.For example, the gamma characteristic of image-region is meant mean value (simple average or weighted average), maximum or the minimum value etc. of the pixel value of the pixel that image-region comprises.
Spatial manipulation mechanism uses the gamma characteristic of ambient image regions, carries out the spatial manipulation of object image area.In the spatial manipulation, with the gamma characteristic weighted average of object image area and ambient image regions.Weight in the weighted average is set according to the difference of the gamma characteristic of object image area and ambient image regions.
In the visual processing apparatus of the present invention, can in the spatial manipulation signal, suppress to be subjected to the influence of the very different image-region of gamma characteristic.For example, even be the image that comprises the border etc. of object, under the situation very different, also can derive suitable spatial manipulation signal with the gamma characteristic of object image area in ambient image regions.Consequently, even in the visual processes of usage space processing signals, also can suppress the generation of false contouring etc.Therefore, can realize a kind of visual processes that improves visual effect.
Visual processing apparatus described in the note 24 is that a kind of wherein, the absolute value of the difference of gamma characteristic is big more according to note 23 described visual processing apparatus, and weighting is just more little.
Here, the dull value that reduces of difference that weight both can be used as according to gamma characteristic provides, and the comparison of difference that again can be by given threshold value and gamma characteristic is set at specified value.
In the visual processing apparatus of the present invention, can in the spatial manipulation signal, suppress to be subjected to the influence of the very different image-region of gamma characteristic.For example, even be the image that comprises the border etc. of object, under the situation very different, also can derive suitable spatial manipulation signal with the gamma characteristic of object image area in ambient image regions.Consequently, even in the visual processes of usage space processing signals, also can suppress the generation of false contouring etc.Therefore, can realize a kind of visual processes that improves visual effect.
Visual processing apparatus described in the note 25 is that a kind of wherein, the distance between object image area and the ambient image regions is big more according to note 23 or 24 described visual processing apparatus, and weighting is just more little.
Here, weight both can be used as according to the dull value that reduces of distance size between object image area and the ambient image regions and had provided, and can be set at specified value by the comparison of given threshold value with the distance size again.
In the visual processing apparatus of the present invention, can be in the spatial manipulation signal, inhibition is subjected to the influence away from the ambient image regions of object image area.For example, even in ambient image regions is the image that comprises the border etc. of object, under the situation very different with the gamma characteristic of object image area, under the situation of ambient image regions away from object image area, can suppress influence, derive more suitable spatial manipulation signal from ambient image regions.
Visual processing apparatus described in the note 26 is a kind of according to each the described visual processing apparatus in the note 23～25, and wherein, image-region is made of a plurality of pixels.The gamma characteristic of object image area and ambient image regions is as the characteristic quantity decision of the pixel value that constitutes each image-region.
In the visual processing apparatus of the present invention, when carrying out the spatial manipulation of each image-region, except the pixel that is comprised in each image-region, also use the gamma characteristic that comprises the pixel that is comprised in the wide area image-region of ambient image regions to handle.Therefore, can carry out more suitable spatial manipulation.Consequently, even in the visual processes of usage space processing signals, also can suppress the generation of false contouring etc.Therefore, can realize a kind of visual processes that improves visual effect.
Visual processing apparatus described in the note 27 has object image area determination means, ambient image regions determination means, greyscale transformation characteristic export agency and gray scale processing mechanism.The object image area determination means, according to the picture signal of being imported, decision becomes the object image area of the derived object of greyscale transformation characteristic.The ambient image regions determination means, decision is positioned at the periphery of object image area, and comprises at least 1 ambient image regions of a plurality of pixels.Greyscale transformation characteristic export agency, the peripheral view data of use ambient image regions, the greyscale transformation characteristic of derived object image-region.The gray scale processing mechanism according to the greyscale transformation characteristic that is derived, carries out the gray scale of the picture signal of object image area and handles.
Object image area for example is by being included in the pixel in the picture signal or picture signal being divided into image block behind the given unit by zone that other a plurality of pixels constituted etc.Ambient image regions for example is that picture signal is divided into other the zone that a plurality of pixel constituted of image block behind the given unit.The periphery view data is the view data of ambient image regions, or from the data that view data derived, for example the pixel value of ambient image regions, gamma characteristic (brightness of each pixel or lightness), thumbnail (downscaled images or reduced resolution between get image) etc.In addition, ambient image regions as long as be positioned at the periphery of object image area, needs not be the zone that surrounds object image area.
In the visual processing apparatus of the present invention, when judging the greyscale transformation characteristic of object image area, use the peripheral view data of ambient image regions to judge.Therefore, handle the effect of adding spatial manipulation can for the gray scale of each object image area, thereby can realize a kind of further gray scale processing that has improved visual effect.
Visual processing method described in the note 28 has object image area deciding step, ambient image regions deciding step, greyscale transformation characteristic derivation step and gray scale treatment step.The object image area deciding step, according to the picture signal of being imported, decision becomes the object image area of the derived object of greyscale transformation characteristic.The ambient image regions deciding step, its decision is positioned at the periphery of object image area, and comprises at least 1 ambient image regions of a plurality of pixels.The greyscale transformation characteristic derives step, uses the peripheral view data of ambient image regions, the greyscale transformation characteristic of derived object image-region.The gray scale treatment step according to the greyscale transformation characteristic that is derived, carries out the gray scale of the picture signal of object image area and handles.
Visual processing method of the present invention when judging the greyscale transformation characteristic of object image area, uses the peripheral view data of ambient image regions to judge.Therefore, handle the effect of adding spatial manipulation can for the gray scale of each object image area, thereby can realize a kind of further gray scale processing that has improved visual effect.
Visual processing program described in the note 29 is that a kind of using a computer carried out the program of visual processing method, and this visual processing method is used for the picture signal of being imported is carried out visual processes.Visual processing method has object image area deciding step, ambient image regions deciding step, greyscale transformation characteristic derivation step and gray scale treatment step.The object image area deciding step, according to the picture signal of being imported, decision becomes the object image area of the derived object of greyscale transformation characteristic.The ambient image regions deciding step, its decision is positioned at the periphery of object image area, and comprises at least 1 ambient image regions of a plurality of pixels.The greyscale transformation characteristic derives step, uses the peripheral view data of ambient image regions, the greyscale transformation characteristic of derived object image-region.The gray scale treatment step according to the greyscale transformation characteristic that is derived, carries out the gray scale of the picture signal of object image area and handles.
Visual processing program of the present invention when judging the greyscale transformation characteristic of object image area, uses the peripheral view data of ambient image regions to judge.Therefore, handle the effect of adding spatial manipulation can for the gray scale of each object image area, thereby can realize a kind of further gray scale processing that has improved visual effect.
Semiconductor device described in the note 30 has object image area determination section, ambient image regions determination section, greyscale transformation characteristic leading-out portion and gray scale handling part.The object image area determination section, according to the picture signal of being imported, decision becomes the object image area of the derived object of greyscale transformation characteristic.The ambient image regions determination section, decision is positioned at the periphery of object image area, and comprises at least 1 ambient image regions of a plurality of pixels.Greyscale transformation characteristic leading-out portion, the peripheral view data of use ambient image regions, the greyscale transformation characteristic of derived object image-region.The gray scale handling part according to the greyscale transformation characteristic that is derived, carries out the gray scale of the picture signal of object image area and handles.
In the semiconductor device of the present invention, when judging the greyscale transformation characteristic of object image area, use the peripheral view data of ambient image regions to judge.Therefore, handle the effect of adding spatial manipulation can for the gray scale of each object image area, thereby can realize a kind of further gray scale processing that has improved visual effect.
(industry application)
Visual processing apparatus among the present invention can be applicable to the picture signal that the gray scale that must realize further improving visual effect is processed is implemented visual processing apparatus that gray scale processes etc.
Claims (3)
1. visual processing apparatus possesses:
Image segmentation portion is divided into a plurality of image-regions with input picture;
Zone conversion characteristics leading-out portion, it is a kind of conversion characteristics leading-out portion, with the wide area image-region lightness export area greyscale transformation characteristic of representing the wide area image-region, above-mentioned wide area image-region, comprise among above-mentioned a plurality of image-region as a plurality of image-regions around the object image area of the object of greyscale transform process, above-mentioned zone greyscale transformation characteristic is at above-mentioned object image area, have above-mentioned wide area image-region lightness big more then with the pixel value conversion of above-mentioned input picture must be more little characteristic; And,
Conversion process portion, the pixel value with the pixel that comprised in the above-mentioned object image area is transformed to according to above-mentioned zone greyscale transformation characteristic and carries out the pixel value that greyscale transform process obtains.
2. visual processing apparatus possesses:
Image segmentation portion is divided into a plurality of image-regions with input picture;
Zone conversion characteristics leading-out portion is a kind of conversion characteristics leading-out portion, with the wide area image-region lightness export area greyscale transformation characteristic of representing the wide area image-region,
Above-mentioned wide area image-region, comprise among above-mentioned a plurality of image-region as a plurality of image-regions around the object image area of the object of greyscale transform process, above-mentioned zone greyscale transformation characteristic is at above-mentioned object image area, have above-mentioned wide area image-region lightness big more then with the pixel value conversion of above-mentioned input picture must be more little characteristic;
The conversion characteristics correction portion according to the locations of pixels coordinate that is comprised in the above-mentioned object image area, is revised above-mentioned zone greyscale transformation characteristic, obtains the pixel grey scale conversion characteristics; And,
Conversion process portion, the pixel value with the pixel that comprised in the above-mentioned object image area is transformed to according to above-mentioned revised pixel grey scale conversion characteristics and carries out the pixel value that greyscale transform process obtains.
3. visual processing apparatus according to claim 2 is characterized in that:
Above-mentioned zone conversion characteristics leading-out portion is also derived the regional conversion characteristics at the image-region around the above-mentioned object image area,
Above-mentioned conversion characteristics correction portion, according to comprising the locations of pixels coordinate in the above-mentioned object image area, to carrying out interior branch, obtain revised pixel grey scale conversion characteristics at the area grayscale conversion characteristics of above-mentioned object image area and at the area grayscale greyscale transformation characteristic of the image-region around the above-mentioned object image area.
Applications Claiming Priority (6)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2003320059 | 2003-09-11 | ||
| JP2003-320059 | 2003-09-11 | ||
| JP2003320059 | 2003-09-11 | ||
| JP2004115166 | 2004-04-09 | ||
| JP2004115166 | 2004-04-09 | ||
| JP2004-115166 | 2004-04-09 |
Related Parent Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2004800262545A Division CN100456800C (en) | 2003-09-11 | 2004-09-10 | Visual processing device, visual processing method, visual processing program, and semiconductor device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101459755A CN101459755A (en) | 2009-06-17 |
| CN101459755B true CN101459755B (en) | 2011-07-13 |
Family
ID=37078443
Family Applications (3)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2004800262545A Active CN100456800C (en) | 2003-09-11 | 2004-09-10 | Visual processing device, visual processing method, visual processing program, and semiconductor device |
| CN2008101843841A Active CN101459755B (en) | 2003-09-11 | 2004-09-10 | Visual processing device |
| CN2008101843856A Active CN101453544B (en) | 2003-09-11 | 2004-09-10 | Gray-scale transformation processing apparatus |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2004800262545A Active CN100456800C (en) | 2003-09-11 | 2004-09-10 | Visual processing device, visual processing method, visual processing program, and semiconductor device |
Family Applications After (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008101843856A Active CN101453544B (en) | 2003-09-11 | 2004-09-10 | Gray-scale transformation processing apparatus |
Country Status (1)
| Country | Link |
|---|---|
| CN (3) | CN100456800C (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5665393B2 (en) | 2010-07-05 | 2015-02-04 | キヤノン株式会社 | Image processing apparatus, image processing method, and program |
| WO2013187133A1 (en) * | 2012-06-12 | 2013-12-19 | 富士フイルム株式会社 | Image processing device and image processing method |
| CN104299187B (en) * | 2013-07-15 | 2018-08-31 | 上海联影医疗科技有限公司 | The image processing method and device of X-ray shooting system |
| CN112288661B (en) * | 2020-12-28 | 2021-04-09 | 成都索贝数码科技股份有限公司 | Image color correction method |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3003561B2 (en) * | 1995-09-25 | 2000-01-31 | 松下電器産業株式会社 | Gradation conversion method and circuit, image display method and apparatus, and image signal conversion apparatus |
| JP3880156B2 (en) * | 1997-10-17 | 2007-02-14 | シャープ株式会社 | Image processing device |
| JP3758452B2 (en) * | 2000-02-28 | 2006-03-22 | コニカミノルタビジネステクノロジーズ株式会社 | RECORDING MEDIUM, IMAGE PROCESSING DEVICE, AND IMAGE PROCESSING METHOD |
| US6473522B1 (en) * | 2000-03-14 | 2002-10-29 | Intel Corporation | Estimating text color and segmentation of images |
| US7116800B2 (en) * | 2001-05-30 | 2006-10-03 | Eaton Corporation | Image segmentation system and method |
-
2004
- 2004-09-10 CN CNB2004800262545A patent/CN100456800C/en active Active
- 2004-09-10 CN CN2008101843841A patent/CN101459755B/en active Active
- 2004-09-10 CN CN2008101843856A patent/CN101453544B/en active Active
Non-Patent Citations (3)
| Title |
|---|
| JP特开2000-4379A 2000.01.07 |
| JP特开2001-243463A 2001.09.07 |
| JP特开平11-122488A 1999.04.30 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101453544A (en) | 2009-06-10 |
| CN101453544B (en) | 2011-03-23 |
| CN1849627A (en) | 2006-10-18 |
| CN100456800C (en) | 2009-01-28 |
| CN101459755A (en) | 2009-06-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102238314B (en) | Visual processing apparatus, visual processing method, program, recording medium, display, and integrated circuit | |
| JP4157592B2 (en) | Visual processing device, display device, visual processing method, program, and integrated circuit | |
| KR101089394B1 (en) | Visual processing device, visual processing method, visual processing program, and semiconductor device | |
| CN100563342C (en) | Image coding method and image decoding method | |
| CN100358366C (en) | Filtering intensity decision method, moving picture encoding method, and moving picture decoding method | |
| CN101242537B (en) | Encoding method | |
| JP4437150B2 (en) | Visual processing device, display device, visual processing method, program, and integrated circuit | |
| CN101686306A (en) | Visual processing device, visual processing method, visual processing program, integrated circuit, display device, imaging device, and mobile information terminal | |
| JP4641784B2 (en) | Gradation conversion processing device, gradation conversion processing method, image display device, television, portable information terminal, camera, integrated circuit, and image processing program | |
| JP5116803B2 (en) | Filter device, filtering method, and filtering program | |
| CN101459755B (en) | Visual processing device | |
| JP4094652B2 (en) | Visual processing device, visual processing method, program, recording medium, display device, and integrated circuit | |
| CN100477721C (en) | Visual processing device, visual processing method, visual processing program, integrated circuit, display device, image-taking device and mobile information terminal | |
| JP4437149B2 (en) | Visual processing device, visual processing method, program, recording medium, display device, and integrated circuit | |
| JP4414464B2 (en) | Visual processing device, visual processing method, visual processing program, and semiconductor device | |
| JP4126298B2 (en) | Visual processing device, visual processing method, visual processing program, and semiconductor device | |
| JP4558409B2 (en) | Filter device, filtering method, and filtering program | |
| JP2005322205A5 (en) | ||
| CN100417175C (en) | Visual processing apparatus, visual processing method, visual processing program, and semiconductor device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| ASS | Succession or assignment of patent right |
Owner name: MATSUSHITA ELECTRIC (AMERICA) INTELLECTUAL PROPERT Free format text: FORMER OWNER: MATSUSHITA ELECTRIC INDUSTRIAL CO, LTD. Effective date: 20140717 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20140717 Address after: California, USA Patentee after: PANASONIC INTELLECTUAL PROPERTY CORPORATION OF AMERICA Address before: Osaka Japan Patentee before: Matsushita Electric Industrial Co.,Ltd. |