JP2000307896A - Device and method for image processing - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image quality improving method and an image processor using the method, which dynamically correct an image signal to reproduce a part desired to be taken in the most attentional way in the image signal, i.e., a part having significant information, without lack of gradation and deficiency of signals. SOLUTION: This device samples a signal what is equivalent to one screen image from a dynamic image signal by every prescribed period and extracts a fixed area from the one sampled image, calculates a statistical characteristic quantity 7 from the statistical distribution of the luminance values of pixels included in the extracted area, decides a luminance value correction parameter 8 according to the statistical characteristic quantity and performs correction conversion 3 of the dynamic image signal with an input/output characteristic applying the luminance value correction parameter.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and an image processing method for displaying a moving image by a digital video signal.

[0002]

2. Description of the Related Art In an image processing apparatus for displaying a moving image using a digital video signal, various attempts have been made to perform various correction processes to improve the display image quality.

As one of the methods, there is a method of improving the image quality of a target object by applying γ (gamma) correction to an image signal so as to expand a dynamic range around a frequently occurring luminance value.

[0004] In addition, a method is also adopted in which the above-mentioned gamma correction characteristics can be switched, and the gamma correction characteristics are selected in accordance with the characteristics of the projected image. For example, in the case of a movie having many night scenes, a correction characteristic for extending / enhancing the vicinity of low luminance is selected. However, in any of the above cases, since the correction characteristics are set in advance, an appropriate correction cannot be obtained according to a changing image.

For example, in an image processing apparatus such as a videophone, the face and the background are the target images to be transmitted and received, but the illuminance of the background and the face changes depending on the other party, the time zone, the lighting condition, and the like. , The face part may be crushed black or fly white. Therefore, automatic and optimal correction with rich gradation is required for the face part, which is the most important information.

[0006]

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above-mentioned problem. In order to reproduce the most noticeable portion and the portion having important information in an image signal without crushing or skipping, the image is reproduced. It is an object to provide an image processing apparatus and an image processing method for dynamically correcting a signal.

[0007]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention provides a sampling means for sampling a picture corresponding to one screen every certain time from a moving picture signal, and a method for sampling one picture sampled by the sampling means. Extracting means for extracting a specific area; means for calculating a statistical feature amount from a statistical distribution of luminance values of pixels included in the area extracted by the extracting means; and a luminance value corresponding to the statistical feature amount. The image processing apparatus further includes means for calculating a value correction parameter, and means for correcting and converting the moving image signal based on input / output characteristics to which the luminance value correction parameter is applied.

That is, the sampling means corresponds to the dynamic correction, the extraction means defines an area for which the image quality is to be improved, and an index of the brightness of the area is obtained by calculating a statistical feature quantity. Since the brightness value correction parameter is calculated and the moving image signal is corrected and converted by applying the brightness value correction parameter, the most noticeable portion of the image signal and the portion having important information are crushed or skipped. Can be reproduced without.

In this case, the input / output characteristics to which the luminance value correction parameter is applied are linear compression / expansion characteristics that are continuous and change in inclination for each input range.

This is for reducing the calculation load of the luminance value correction parameter and the hardware scale for performing the correction conversion.

Further, the statistical distribution is a frequency distribution (histogram) for each luminance value range or a distribution function thereof.

This is to reduce the burden of calculating the statistical feature amount.

Here, the statistical feature value is an average value.

This is to reduce the burden of calculating the statistical feature.

Here, the statistical feature amount is a ratio equal to or less than a certain luminance value.

This is to reduce the burden of calculating a statistical feature.

Here, the statistical feature amount is a luminance value when the distribution function reaches a certain value.

This is to reduce the burden of calculating the statistical feature. Further, since the range to be expanded in the input / output characteristics of the correction conversion can be set independently of the upper limit and the lower limit, more appropriate correction conversion characteristics can be obtained.

The input / output characteristic to which the luminance value correction parameter is applied is one of a plurality of input / output characteristic patterns prepared in advance, and the means for performing the correction conversion is: The moving image signal is corrected and converted by the input / output characteristic which changes with a hysteresis characteristic by applying the luminance value correction parameter.

This is the hardware for correction conversion,
This is in order to significantly reduce the size of the software. Also,
Since switching is performed with the hysteresis characteristic, it is possible to prevent side effects due to correction conversion such as flickering of the display screen.

Further, instead of switching based on the hysteresis characteristic, switching may be performed only when the state to be switched is repeated a plurality of times. Also in this case, it is possible to prevent side effects due to correction conversion such as flickering of the display screen.

Further, one screen equivalent is sampled at a certain time interval from the moving image signal, a specific area is extracted from the sampled one screen, and the specific area is extracted from the statistical distribution of the luminance values of the pixels included in the extracted area. Calculate statistical features,
The method is characterized in that a brightness value correction parameter is calculated according to the statistical feature amount, and the moving image signal is corrected and converted according to input / output characteristics to which the brightness value correction parameter is applied.

That is, the sampling corresponds to the dynamic correction, the extraction defines an area in which the image quality is to be improved, and the statistical feature is calculated to obtain an index of the brightness of the area. Correction parameters are calculated, and the moving image signal is corrected and converted by applying the luminance value correction parameter. Therefore, the most noticeable part of the image signal and the part having important information are reproduced without crushing or skipping. it can.

[0024]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing a configuration of an image processing apparatus according to a first embodiment of the present invention. In the present embodiment, the present invention is applied to a device such as a videophone that performs processing up to displaying a digital video signal of the other party transmitted through a line or a local digital video signal captured by a video camera. It is.

As shown in FIG.
Transmission / reception unit 1, image capture unit 2, correction conversion unit 3, image display unit 4, user setting unit 5, screen sampling and constant region extraction unit 6, statistical feature amount calculation unit 7, calculation of luminance value correction parameters And a display device 9.

The transmission / reception unit 1 receives a digital video signal transmitted through a line, and sends a digital video signal of a local image captured by an image capturing unit 2 such as a video camera to the line. The digital video signal received by the transmission / reception unit and the digital video signal captured by the image capture unit 2 are guided to the correction conversion unit 3 and the screen sampling and fixed area extraction unit 6 of the present apparatus.

Screen sampling and fixed area extracting section 6
Performs a process of sampling a signal corresponding to one screen every certain time from a moving image signal from the transmitting / receiving unit 6, extracting a specific area from the sampled one screen, and outputting the extracted area to the statistical feature calculating unit 7. Is what you do.

The statistical feature calculating unit 7 calculates the statistical feature from the statistical distribution of the luminance values of the pixels included in the area extracted by the screen sampling and the constant area extracting unit 6.

The luminance value correction parameter calculation unit 8 calculates a luminance value correction parameter from the statistical feature value calculated by the statistical feature value calculation unit 7.

The correction conversion section 3 corrects the digital video signal supplied from the transmission / reception section 1 by using the luminance value correction parameters obtained by the luminance value correction parameter calculation section 8. The corrected digital video signal is converted by the image display unit 4 into a signal that can drive the display device 9.

The user setting section 5 is provided for the screen sampling and constant area extracting section 6 and serves as a user of correction parameters.
The setting can be made.

Next, the operation of the image processing apparatus of this embodiment will be described.

A moving image to be transmitted is captured by the image capturing unit 2, and the captured moving image is transmitted to the transmitting / receiving unit 1.
Sent to the line via

On the other hand, the moving image from the other party received by the transmission / reception unit 1 from the line and the local moving image captured by the image capturing unit 2 are sent to the correction conversion unit 3, the screen sampling unit and the fixed area extracting unit 6. Is output.

The moving image signal input to the correction converter 3 is
Correction conversion processing is performed. The correction conversion method converts an input luminance value of each pixel into an output luminance value according to a fixed input / output characteristic. The corrected and converted moving image signal is output to the image display unit 4 that drives the display device 9, and the image display unit 4 displays the received moving image on the display device 9.

The moving image signal input to the screen sampling and fixed area extracting section 6 is sampled at regular intervals. The sampling time interval is one of the correction parameters that determines how finely the characteristic change of the correction conversion is performed in time, and a predetermined time is set in advance or the user can set the time. It depends on the setting. In the case of the user setting, the user sets the user setting section 5.

Generally, if this time interval is set short,
Appropriate correction conversion can be realized depending on the state of the transmitted moving image, but on the other hand, there is a possibility that it may react to noise or the like, but may cause image quality deterioration such as flicker. On the other hand, if the time interval is set to be long, phenomena that are completely opposite to these occur, so this setting is an intermediate value depending on the noise quality of the received signal and the relationship between this correction conversion and other parameter settings. It is preferable to select and set appropriate values.

Screen sampling and fixed area extracting section 6
Extracts a certain area from the sampled screen.

The certain area is an area which is most noticeable and has information important to the receiver. In other words, in the case of a videophone, it is the face of the other party. The extraction of the face part can be done in a faithful way, such as by contour extraction, or in the nature of a video phone, by using the fact that the other person's face is almost at the center of the screen, making it a fixed central area of the screen It is. In any case, it is possible to eliminate the influence of the background and obtain a display with a rich gradation of the face of the other party.

In the case of contour extraction, the correction parameters are calculated from the information of the entire face of the partner, so that the appropriateness of the correction is improved. However, on the other hand, hardware and software are increased. The opposite is true when a central fixed area is used. Therefore, it is preferable to make a selection according to what cost and how much correction effect is expected.

The output of the screen sampling and constant area extracting unit 6 is led to a statistical characteristic amount calculating unit 7 for calculating the statistical characteristic amount of the luminance value of the pixel belonging to the extracted constant area. .

Each pixel belonging to a certain area has information on the brightness to be displayed, that is, a luminance value. Therefore, it is possible to create a frequency distribution (histogram) with the luminance value on the horizontal axis and the number of pixels on the vertical axis.

Here, the input pixels may be pixels that belong to the above-mentioned fixed area and are sampled at appropriate intervals (for example, one sample for every two pixels, one sample for every four pixels, etc.). It is.

When each pixel is used, all information is used, so that the appropriateness of correction is improved.
This leads to an increase in feature calculation hardware and software,
On the other hand, the longer the sample interval, the opposite is true. Therefore, it is preferable to make a selection according to what cost and how much correction effect is expected. It should be noted that the user setting unit 5 can select at what interval sampling is performed.

Also, for the luminance value of each pixel, a method of counting all the bits represented by digital values and counting the number of frequencies, and a method of counting the number of frequencies according to some upper bits, that is, a certain range width. There is a way to do it.

If the frequency is counted by setting the input range width more finely, more information is used, so that the appropriateness of the correction is improved, but on the other hand, hardware for calculating the feature amount and software are increased. Therefore, it is preferable to set the input range width depending on how much cost and how much correction effect is expected.

By the way, the above frequency distribution can be made a distribution function by accumulating the number of pixels below a certain input luminance value.

The statistical feature can be calculated from the frequency distribution and the distribution function. The statistical feature amount includes, for example, an average value of input luminance values, a ratio of a certain input luminance value or less, an input luminance value when a distribution function reaches a certain value, and the like.

These values serve as indices indicating whether the region to which the sampled pixel belongs is bright or dark as a whole. For example, a small average value, a large ratio below a certain input luminance value, a small distribution input function when the input luminance value reaches a certain value, etc. indicate that the area is dark, and vice versa. Indicates that the area is bright.

Next, the statistical feature quantity which is an index of the brightness of these areas is provided to the brightness value correction parameter calculating section 8.

The luminance value correction parameter is a parameter for specifying the input / output characteristics of the correction converter 3. Therefore, the parameter needs to specify a function expressing the input / output characteristics. For example, if the function is a linear function (straight line), the intercept and the slope are quadratic functions. If the function is a quadratic function, there are an intercept, a primary coefficient, a secondary coefficient, and the like. If applicable, these information for each category.

Qualitatively, since the brightness in the area can be known from the statistical feature amount, the above function is set so as to extend the dark input range when dark, and to extend the bright input range when bright. Set the above function as follows. For example, when the function is a continuous linear function (straight line), that is, a polygonal line, when the statistical feature indicates darkness, the slope of the function in the low range of the input luminance value is set to be steep, and conversely, the function is set to be bright. When the parameter is indicated, a parameter for steepening the slope of the function in a range where the input luminance value is high is calculated.

Here, the steep slope of the function means that
In the correction conversion unit 3, the change range of the input luminance value in the input range is expanded in the output luminance value, which means that the image has more gradation.

By providing the parameters calculated in this way to the correction conversion unit 3, the correction conversion unit 3 has a function for correction, that is, an input / output characteristic specified.

As described above, the input / output characteristic is a characteristic in which the gradation is to be expanded in a certain area, that is, a notable area and an area having important information. Therefore, such an area can be reproduced without crushing or skipping.

Next, a second embodiment will be described with reference to FIGS.

FIG. 4 is a graph showing the characteristics of the correction converter 3, that is, the input / output characteristics in which the input luminance value is shown on the horizontal axis and the output luminance value is shown on the vertical axis. As shown in the figure, the input / output characteristic as the correction conversion characteristic is a linear compression / expansion characteristic that is continuous and changes in slope for each input range.

That is, the input is linearly compressed in the range of 0 to 42a, linearly expanded in the range of 42a to 42b, and linearly compressed in the range of 42b or more. You. With respect to these, the degree of compression and expansion can be determined by comparing with the linear characteristics indicated by a broken line obliquely from the origin.

Of these characteristics, 42a value to 42b
The range represented by the value is a portion corresponding to a luminance value frequently appearing in a certain region to be noted. This setting corresponds to the brightness of this certain region indicated by the statistical feature amount. Are both large, and if dark, 42a, 42
b is both small. As a result, the dynamic range is expanded for the certain area, so that an image without collapse or skipping can be obtained.

Further, by providing such continuous linear characteristics, the number of parameters defining the function can be minimized, and the burden on both parameter calculation and correction conversion is reduced. be able to.

The parameters defining the input / output characteristics can be uniquely given by, for example, the intercept 41a, the slope 40a, the range 42a, the slope 40b, the range 42b, and the slope 40c shown in FIG. For these calculations, the values of 41b and 41c can be used if necessary. That is, the above parameters are calculated from the statistical feature amount obtained by the statistical feature amount calculation unit 7 and provided to the correction conversion unit 3 as correction conversion characteristics, whereby the input / output characteristics of the correction conversion unit 3 are defined. .

As is apparent from FIG. 4, the input / output conversion of the correction conversion unit 3 is performed by adding 41a + 40 from 0 to 42a.
a value × input luminance value, and from 42a value to 42b value, 41b value + 40b value × (input luminance value−42a value), and above 42b value, 41c value + 40c value × (input luminance value−42b value) Become.

It should be noted that these values are, in an actual device, each of the values 40a, 40b, 40c 2 bits, 41 bits.
When each value of a, 41b, 41c and each value of 42a, 42b are expressed by 8 bits, a good balance is achieved in realizing necessary correction characteristics.

Further, the continuous linear characteristic is not limited to three straight lines as shown in FIG. 4, but it is also possible to make a linear approximation closer to an S-shaped curve by using a larger number of straight lines.

When a large number of straight lines are used, the degree of compression / expansion can be set more finely, so that the arbitrariness of correction is improved, such as making the unnaturalness of the boundary of the correction characteristic less noticeable. -Increases hardware and software. Therefore, it is preferable to make a selection based on how much cost and how much arbitrariness and effect are expected.

Next, a third embodiment will be described.

As described above, the brightness value of each pixel belonging to a certain area received by the statistical feature calculator 7 from the screen sampling and certain area extracting unit 6 in FIG.
When counting the frequency, there are a method of counting all the bits represented by the digital value and counting the frequency, and a method of counting the frequency of the partition by some upper bits, that is, by a certain range width.

If the frequency is counted by setting the input range width more finely, more information is used, so that the appropriateness of the correction is improved. On the other hand, however, the hardware for calculating the feature amount and the software are increased. In other words, there is a case where there is a cost merit in creating a frequency distribution or its distribution function for some upper bits, that is, for each range. Therefore, in this embodiment, a frequency distribution for each luminance value range or a distribution function thereof is used as a statistical distribution.

Next, a fourth embodiment will be described with reference to FIGS.

6 (a), 6 (b) and 6 (c) show examples 31 of the frequency distribution in the statistical feature quantity calculation unit 7,
It is a figure which shows 32, 33, (a) when a fixed area | region is bright, (b) when a fixed area | region is a medium brightness, (c)
Is a case where a certain area is dark. These frequency distributions 31,
As shown in FIG. 6, an average value of the input luminance values is calculated from 32 and 33 as a statistical feature amount.

The average value is sent to the luminance value correction parameter calculating section 8, and the luminance value parameter calculating section 8 calculates the input luminance value range to be expanded. This calculation is performed in a certain range including the average value as the center. This is because it can be seen that the luminance value near the average value is the most frequent, and is near the frequent luminance value in a certain area.

That is, the input / output characteristics of the correction converter 3 given these parameters are shown in FIG.
It becomes like 61,62,63 of three figures under (b) and (c). As is clear from the upper and lower figures, the correction conversion unit 3 has expanded input / output characteristics in a range where the input luminance value frequently appears. Therefore,
It is possible to obtain a display with rich gradation in a certain area having important information to be noted.

The characteristics of the correction conversion unit 3 are continuously changed by the calculated average value including the cases of the three examples 61, 62 and 63, and the screen sampled by the screen sampling is updated. It can change every moment.

Next, a fifth embodiment will be described with reference to FIG. 7 together with FIG.

The upper three figures in FIGS. 7A, 7B, and 7C show examples 31 of the frequency distribution in the statistical feature calculator 7.
It is a figure which shows 32, 33, (a) when a fixed area | region is bright, (b) when a fixed area | region is a medium brightness, (c)
Is a case where a certain area is dark. These frequency distributions 31,
From 32 and 33, the ratio of pixels occupying 77 values, 78 values, and 79 values or less, which are constant and the same value, is calculated as the statistical feature amount.

That is, the frequency distribution 31 is calculated for all the pixels 74a +
Since the pixels composed of 74b and the pixels having a value of 77 or less are the portion 74a indicated by oblique lines, the ratio is expressed by the number of 74a pixels / (7
4a pixel number + 74b pixel number). The same applies to the case of medium brightness (b) and darkness (c).
a, 75b, 76a, and 76b.

As is apparent from FIG. 7, the calculated value is an index of whether the certain area is bright or dark.

Then, the calculated value is sent to the luminance value correction parameter calculating section 8, and the luminance value parameter calculating section 8 calculates the input luminance value range to be expanded. This calculation is performed so that the smaller the value, the higher the level is expanded. This is because, as the calculated value is smaller, it can be seen that a larger luminance value frequently appears, and the vicinity of the frequently appearing luminance value is expanded. That is, the input / output characteristics of the correction conversion unit 3 given these parameters are as shown in 71, 72, and 73 in three figures below FIGS. 7A, 7B, and 7C, respectively. As is clear from the upper and lower figures, the correction conversion unit 3 has expanded input / output characteristics in a range where the input luminance value frequently appears. Therefore, a display with a rich gradation can be obtained for a certain area having important information to be noted.

The characteristics of the correction conversion unit 3 are continuously changed by the calculated values, including the cases of the three examples 71, 72 and 73, and the screen sampled by the screen sampling is updated. It can change every moment.

The above 77 values, 78 values and 79 values are
In addition to the fixed value, it is also possible to allow the user setting section 5 to select and set the value.

Further, another example will be described.

The above-described calculation of the ratio of the number of pixels below a certain luminance value, which is a statistical feature, is performed in the same manner.
This is compared with a plurality of threshold values to determine which rank is located. As a result, one of a plurality of input / output characteristic patterns prepared in advance and prepared by the correction conversion unit 3 is selected.

That is, this corresponds to a case where the number of conversion pallets for correction conversion by the correction conversion unit 3 is limited. In this case, the burden on the hardware and software of the correction conversion unit 3 can be reduced.

In this case, FIG. 13 shows, for reference, the flow of operation when a fixed area is set as a fixed area of the screen. That is, steps 131 to 13 of FIG.
It looks like 5.

Next, a sixth embodiment will be described with reference to FIG. 8 together with FIG.

8 (a), 8 (b) and 8 (c) show examples 31 of the frequency distribution in the statistical feature calculating unit 7, and FIG.
It is a figure which shows 32, 33, (a) when a fixed area | region is bright, (b) when a fixed area | region is a medium brightness, (c)
Is a case where a certain area is dark. These frequency distributions 31,
The input luminance value when the frequency distribution function reaches a certain value as shown in FIG. In this example, one value is 50%.

The calculated value is sent to the luminance value correction parameter calculating section 8, and the luminance value parameter calculating section 8 calculates the input luminance value range to be expanded. This calculation is performed in a certain range including the calculated value substantially at the center. This is because it can be seen that the luminance value near the calculated value is the most frequent, and the luminance value is near the frequent luminance value in a certain area.

That is, the input / output characteristics of the correction converter 3 given these parameters are shown in FIG.
It becomes like 81,82,83 of three figures under (b) and (c). As is clear from the upper and lower figures, the correction conversion unit 3 has expanded input / output characteristics in a range where the input luminance value frequently appears. Therefore,
It is possible to obtain a display with rich gradation in a certain area having important information to be noted.

It should be noted that the characteristics of the correction conversion unit 3 change continuously including the cases of the above three examples 81, 82 and 83 according to the calculated values, and every time a screen sampled by screen sampling is updated. Can change.

In addition to the above-mentioned certain value being a fixed value such as 50%, the value can be selected and set from the user setting section 5.

Next, another example will be described with reference to FIG. 3 together with FIG.

The upper three figures in FIGS. 3A, 3B and 3C show examples 31 of the frequency distribution in the statistical feature quantity calculator 7;
It is a figure which shows 32, 33, (a) when a fixed area | region is bright, (b) when a fixed area | region is a medium brightness, (c)
Is a case where a certain area is dark. These frequency distributions 31,
The input luminance value when the frequency distribution function reaches a certain value as the statistical feature amount from 32 and 33 is calculated. In this example, there are two values, 3% and 97%, and the input luminance value is calculated for each of them.

The calculated value is sent to the luminance value correction parameter calculating section 8, and the luminance value parameter calculating section 8 calculates the input luminance value range to be expanded. This calculation is based on the lower limit of the 3% value point,
This is done with the 97% value point as the upper limit. This is because it can be seen that the luminance value between these two calculated values is the most frequent, and the luminance value is in a range around the frequent luminance value of a certain area to be noted.

That is, the input / output characteristics of the correction converter 3 given these parameters are shown in FIG.
(B) and (c) below.

As is clear from the upper and lower figures, the range 34 in which the input luminance value frequently appears is clear.
b, 35b, and 36b, the correction conversion unit 3 has expanded input / output characteristics, and the other ranges 34a, 35a, 3b
The input / output characteristics are compressed in 6a, 34c, 35c, and 36c. Therefore, a display with a rich gradation can be obtained for a certain area having important information to be noted.

Also, in this method, since the upper and lower limits to be expanded are set independently, there is a feature that the range to be expanded can be appropriately set according to the image as compared with the above-described embodiment.

Note that the characteristics of the correction conversion unit 3 continuously change in accordance with the calculated values, including the above three examples, and may change every time a screen sampled by screen sampling is updated.

The above-mentioned certain value may be set to a fixed value such as 3% or 97%, or may be selectively set by the user setting section 5.

Next, a seventh embodiment will be described with reference to FIGS. 9 to 12 together with FIG. In the above description, the input / output characteristics of the correction conversion unit 3 have been described mainly on the case where the input / output characteristics change continuously due to the change of the provided parameters. In this case, the conversion pallets to be prepared by the correction conversion unit 3 need only correspond to the values of the parameters that can be changed.

Therefore, the equipment burden of the hardware and software becomes large.

Therefore, in the present embodiment, the correction conversion unit 3
Are a plurality of (for example, four) input / output characteristic patterns prepared in order and prepared in advance.

FIG. 9 shows an example of input / output characteristics in these four cases.
Shown in In the order of (a), (b), (c), and (d), the input / output characteristic pattern to be applied when a certain area has a bright luminance value.

That is, in the order of the characteristics 91, 92, 93, and 94, the characteristics extend from a bright range to a dark range.

As a result, the burden on hardware and software can be greatly reduced.

For example, if the number of pixels in a certain area is 172 × 14
4, 24,768 pixels, and the ratio of the number of pixels below a certain reference value is used as the statistical feature amount.
Corresponds to the selection characteristic of the conversion palette of the correction conversion unit 3 as shown in FIG. As a result, it is possible to perform correction conversion according to the luminance value of a certain area although the texture is coarse.

This roughness may cause a side effect when the number of pixels below a certain reference value causes a temporal change such as 120 shown in FIG. That is, the switching value 18
This is because, when the number of pixels is calculated over time around 576 pixels, the selected input / output characteristic pattern frequently changes over time, which causes flickering of the display screen. This is the same before and after other switching values.

Therefore, in this embodiment, switching of the input / output characteristic pattern is performed with a hysteresis characteristic.

For example, if the number of pixels is the same as the above example,
A correction conversion characteristic as shown in FIG. 11 is selected. That is, when the number of pixels having a calculated brightness value or less passes through 7076 pixels, 14152 pixels, and 21228 pixels, the selection changes as shown in FIG. The number of pixels below the value is 17690 pixels, 1
When going down after passing through 0614 pixels and 3538 pixels, the selection changes as shown in FIG.

Thus, even if the number of pixels below a certain luminance value is calculated over time before or after the number of pixels whose selection changes, the selected correction conversion characteristic does not change frequently, so that flicker on the screen is prevented. it can.

In addition to the method of giving the hysteresis characteristic to the switching of the input / output characteristic pattern, the number of pixels below a certain luminance value to be calculated is the number of pixels to be continuously switched in several screens to be sampled. Only when the condition becomes, the method of switching the input / output characteristic pattern can be adopted.

Next, another example will be described with reference to FIG. 2 together with FIG.

The moving picture from the other party guided from the transmitting / receiving section 1 to the screen sampling and constant area extracting section 6 changes every moment as shown at 20a to 20e in FIG. From this, the screen is sampled at regular intervals and a sample screen 21 is taken out.

In this embodiment, a face area 22 is extracted from the sample screen 21 as a fixed area. This is because the face area 22 is the area that is most noticeable to the recipient and has important information.

The extraction of the face area requires the extraction of its outline. For this purpose, various outline extraction methods can be used. Since various contour extraction methods can be applied, details thereof are omitted here.

Next, another example will be described with reference to FIG. 5 together with FIG.

In the above case, the face area is extracted as the fixed area. In this embodiment, the fixed central area 51 of the sample screen 21 is set as the fixed area. This is because, in a device such as a videophone, the face of the most notable partner is often located near the center of the screen.

As a result, it is possible to eliminate the burden of hardware and software such as the extraction of the face area, and to obtain a great cost advantage.

The above-mentioned fixed area may be a fixed fixed area, or may be selectively set by the user setting section 5.

[0120]

As described above in detail, according to the first and eighth aspects of the present invention, the gradation characteristic is expanded for the most noticeable portion of the image signal and the portion having important information. Therefore, an image processing apparatus and an image processing method for dynamically correcting an image signal without crushing or skipping can be realized.

According to the second aspect of the present invention, the input / output characteristic to which the luminance value correction parameter is applied is a linear compression / expansion characteristic in which the inclination is continuous and the slope changes for each input range. Therefore, it is possible to reduce the calculation load of the luminance value correction parameter and the scale of hardware for performing the correction conversion.

According to the third aspect of the present invention, the statistical distribution is a frequency distribution (histogram) for each luminance value range or a distribution function thereof. it can.

Further, according to the present invention, since the statistical feature value is an average value, the burden of calculating the statistical feature value can be reduced.

According to the fifth aspect of the present invention, since the statistical feature amount is a ratio equal to or less than a certain luminance value, the burden of calculating the statistical feature amount can be reduced.

According to the present invention, since the statistical feature is a luminance value when the distribution function reaches a certain value, it is possible to reduce the burden of calculating the statistical feature. Further, since the range to be expanded in the input / output characteristics of the correction conversion can be set independently of the upper limit and the lower limit, more appropriate correction conversion characteristics can be obtained.

According to the seventh aspect of the present invention, the input / output characteristics to which the luminance value correction parameter is applied include a plurality of input / output characteristic patterns prepared in advance.
The correction conversion means corrects and converts the moving image signal by the input / output characteristics which change with hysteresis characteristics by applying the luminance value correction parameter, so that hardware for correction conversion is used. , Can significantly reduce the size of the software. Further, since the switching is performed with the hysteresis characteristic, it is possible to prevent side effects due to correction conversion such as flickering of the display screen.

Further, instead of switching with the hysteresis characteristic, switching is performed only when the state to be switched is repeated a plurality of times.
Also in this case, it is possible to prevent side effects due to correction conversion such as flickering of the display screen.

[Brief description of the drawings]

FIG. 1 is a diagram showing a first embodiment of the present invention.

FIG. 2 is an explanatory diagram of a function example of a screen sampling and fixed area extracting unit 6;

FIG. 3 is an explanatory view of a sixth embodiment of the present invention.

FIG. 4 is an explanatory view of a second embodiment of the present invention.

FIG. 5 is an explanatory diagram of another example of the function of the screen sampling and fixed area extraction unit 6;

FIG. 6 is an explanatory diagram of a fourth embodiment of the present invention.

FIG. 7 is an explanatory view of a fifth embodiment of the present invention.

FIG. 8 is an explanatory view of another sixth embodiment of the present invention.

FIG. 9 is an explanatory view of a seventh embodiment of the present invention.

FIG. 10 is an explanatory diagram of a seventh embodiment of the present invention.

FIG. 11 is an explanatory view of a seventh embodiment of the present invention.

FIG. 12 is an explanatory view of a seventh embodiment of the present invention.

FIG. 13 is a flowchart of a process according to another fifth embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Transmission / reception part 2 Image capture part 3 Correction conversion part 4 Image display part 5 User setting part 6 Screen sampling and fixed area extraction part 7 Statistical feature amount calculation part 8 Luminance value correction parameter calculation part

Claims (8)

[Claims] 1. A sampling means for sampling one screen equivalent every certain time from a moving image signal, an extracting means for extracting a specific area from one screen sampled by the sampling means, Means for calculating a statistical characteristic amount from a statistical distribution of luminance values of pixels included in the region; means for calculating a luminance value correction parameter according to the statistical characteristic amount; and applying the luminance value correction parameter. Means for correcting and converting the moving image signal according to the input / output characteristics. 2. The image processing apparatus according to claim 1, wherein the input / output characteristic to which the luminance value correction parameter is applied is a linear compression / expansion characteristic in which a slope changes continuously for each input range. 3. The image processing device according to claim 1, wherein the statistical distribution is a frequency distribution (histogram) for each luminance value range or a distribution function thereof. 4. The image processing apparatus according to claim 1, wherein the statistical feature amount is an average value. 5. The image processing apparatus according to claim 1, wherein the statistical feature amount is a ratio equal to or less than a certain luminance value. 6. The image processing apparatus according to claim 1, wherein the statistical feature amount is a luminance value when the distribution function reaches a certain value. 7. The input / output characteristic to which the luminance value correction parameter is applied is one of a plurality of input / output characteristic patterns prepared in advance, and the means for performing the correction conversion includes: 2. The image processing apparatus according to claim 1, wherein said moving image signal is corrected and converted by said input / output characteristic which changes with a hysteresis characteristic by applying a value correction parameter. 8. A screen corresponding to one screen is sampled at a certain time interval from a moving image signal, a specific area is extracted from the sampled one screen, and a specific area is extracted from a statistical distribution of luminance values of pixels included in the extracted area. Calculating a statistical feature amount, calculating a brightness value correction parameter according to the statistical feature amount, and correcting and converting the moving image signal based on input / output characteristics to which the brightness value correction parameter is applied. Image processing method.