JPH09247483A - Picture emphasizing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a picture emphasizing device where a picture with strong contrast is obtained by deciding the conversion quantity of a density, a hue and a saturation in an output picture. SOLUTION: An HLC signal being the three-attribute signal of visual sensation is obtained in a coordinate converting part 1 and the HLC signal is inputted to a histogram count part 2. A two-dimensional or one-dimensional histogram obtained in the histogram count part 2 is inputted to a color area dividing part 3. The color area is divided by the inputted two-dimensional or one- dimensional histogram in the color area dividing part 3. The divided area is inputted to a shortest distance locating part 4. Distance, that is, to which color area the picture element of a picture signal is close is judged and the nearest area is judged. One look-up table(LUT) for conversion is selected in accordance with the judged area and conversion for emphasizing the density, the saturation and the hue is executed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image enhancing device applicable to, for example, a color copying machine, a color printer device or the like.

[0002]

2. Description of the Related Art For example, an image forming apparatus shown in FIG. 9 is known as an image forming apparatus used in a color copying machine or the like. In FIG. 9, the image forming apparatus includes a scanner unit 31,
The filter unit 32, the color conversion unit 33, the scaling / processing unit 34, the gradation conversion unit 35, and the printer unit 36 are formed. The scanner unit 31 reads a document and obtains image data. The image data is subjected to various kinds of processing every time it passes through the filter unit 32, the color conversion unit 33, the scaling / processing unit 34, and the gradation conversion unit 35. The processed image data is printed on the transfer paper by the printer unit.

[0003]

When printing is performed using the image forming apparatus having the above-described structure, the density distribution of each image is biased, so that the density range of the output device can be effectively utilized. Instead, an image with low contrast is printed.

An image forming apparatus such as a color copying machine or the above-mentioned color printer apparatus, as shown in FIG.
Alternatively, there is a color range that cannot be reproduced because the color reproduction range is narrower than that of the image on the display. Therefore, the optimum contrast cannot be obtained only by printing the read image or the image on the display as it is.

Therefore, in order to obtain the optimum gradation characteristics, statistical data such as density frequency and frequency distribution is used as information obtained from the read image or the image on the display.

As a method for obtaining the optimum gradation characteristics, there are disclosed in Japanese Patent Laid-Open No. 2-050858 and Japanese Patent Laid-Open No. 3-1.
7 1973, JP-A-3-120958, and JP-A-5-075222 are known.

In the printing apparatus described in Japanese Patent Laid-Open No. 2-050858, each gradation data of the output image is converted based on the gradation value obtained by averaging the variation of the frequency for each gradation value, and the image of the number of pixels is displayed. -The object of the present invention is to make it possible to print an easy-to-read image in which light and shade are averaged by printing the content developed in the jig buffer according to a predetermined printing method. In order to achieve such an object, when the contrast improving process is executed, the grayscale frequencies are averaged with 32 grayscale levels based on the average frequency 280.4. If the variation in frequency by gradation is small, the contrast is averaged from white (gradation value 0) to black (gradation value 31), and an easy-to-read image in which shades are averaged is printed. Therefore, when the print key is input, the gradation print data is retransmitted from the host computer, and the gradation value specified by each gradation data is converted into the converted table according to the conversion table. Converted to key value. Then, the gradation pattern of the scale value is read out from a preset gradation table of 32 gradations and is developed by a bit image in the image buffer, and is printed by the printing unit in a predetermined printing format. The configuration is such that an easy-to-see image with averaged contrast is printed.

Further, the gradation image reading apparatus disclosed in Japanese Patent Laid-Open No. 3-171973 changes the gradation correction method in accordance with the frequency of appearance of each density level of the original, thereby improving the gradation of the original. The purpose is to obtain high-contrast gradation from a low-contrast original without damaging it. To achieve these goals,
The minimum density appearing in the manuscript is Dmin, and the maximum density is Dm.
If it is ax, the CPU detects the Dmin and Dmax by referring to the histogram memory. When D> Dmin, D ′ = 0, Dmin <D <Dmax
In case of D ′ = F {255 / (Dmax-Dmin) *
(D-Dmin)}, when Dmax <D, D ′ = 255
The γ correction table is calculated based on Next, the switch is switched to the γ correction unit side to perform scanning for outputting an image signal. With such a configuration, it is possible to always obtain a high-contrast reproduced image from a low-contrast original image.

Further, the image reading apparatus described in Japanese Patent Laid-Open No. 3-120958 impairs the gradation of the original by expanding the range of density levels appearing at a predetermined frequency or more so as to take a predetermined total density level. The purpose is to obtain a high-contrast gradation from a low-contrast original. In order to achieve such an object, the adder as the appearance frequency detecting means detects the appearance frequency (density histogram) of each density level, and the gradation processing section as the expanding means is stored in the dither pattern memory. Based on the stored dither pattern, gradation processing is performed so as to take all predetermined density levels. That is, the density of the original image is quantized by the quantizing means, the appearance frequency of each density level quantized by the quantizing means is detected by the appearance frequency detecting means, and the predetermined frequency among the appearance frequencies detected by the appearance frequency detecting means is detected. The range of density levels appearing above is expanded by the expansion means so as to have a predetermined total density level. This makes it possible to obtain a high-contrast image from a low-contrast original.

Further, the gradation conversion method and apparatus described in JP-A-5-075222 includes a standard gradation conversion curve,
By performing gradation conversion that incorporates the characteristics of the density distribution state that reflects the characteristics of the original image, it is possible to perform gradation conversion according to the characteristics of the original image, based on the characteristics of general gradation conversion. Has an aim. In order to achieve such an object, the image signal after the shading correction is passed through a switch circuit and a histogram counting circuit and a look-up taper.
It is configured such that it can be selectively applied to one of the gradation correction circuits including a bull-type RAM. The circuit obtains a cumulative density histogram from the density distribution in the original image, determines the gradation conversion table based on the result, and stores it in the RAM. The memory stores data expressing a general standard gradation conversion curve. For this reason, a natural reproduction image is ensured by the general gradation conversion characteristic of the standard gradation conversion curve, and then the characteristics of the original image are incorporated into the gradation conversion curve based on the density distribution state of the original image. As a result, it is possible to perform optimal gradation conversion by combining these two properties in a well-balanced manner.

However, as described in JP-B-5-075222, it is disadvantageous in terms of cost to have a large number of standard density curves, and moreover, the margin remaining amount is large each time an input is made. However, there is a problem in that a device that processes in real time, such as a copying machine or a printer device, is disadvantageous in terms of speed.

Further, Japanese Patent Laid-Open No. 2-050858,
JP-A-3-171973, JP-A-2-12095
The one described in Japanese Patent Publication No. 8 has a simple structure and is advantageous in terms of calculation speed and memory capacity. However, since the entire image is contained in one density range, the expression is unreasonable. Actually, in plate making, in order to save it, the gradation characteristics are optimally adjusted and output for each pattern of interest.

The present invention has been made in order to solve the above problems of the prior art, and an object of the present invention is to provide an image enhancing apparatus capable of obtaining an image with high contrast.

[0014]

According to the first aspect of the present invention,
A coordinate transformation unit that obtains at least two-dimensional or more data in which the color space of the image data expresses a color, a region dividing unit that classifies into a plurality of regions based on the appearance frequency distribution of the two-dimensional or more data, and the density of the output image. , An image conversion unit that determines the conversion amount of hue and saturation.

According to a second aspect of the invention, the image conversion unit is R
It is a two-dimensional color space of hue and saturation that is converted from a data value of the GB color space into an L * a * b * color space and further determined from the a * b * plane, or a one-dimensional color space of hue. It is characterized by

According to a third aspect of the invention, the image conversion unit is R
It is characterized in that it is a two-dimensional color space of hue and saturation directly calculated from data values of the GB color space or a one-dimensional hue space of hue.

According to a fourth aspect of the present invention, the area dividing section divides the two-dimensional color data frequency distribution indicating the tint of the input image or the one-dimensional hue frequency distribution into an integral number of color space areas. It is characterized by being.

According to a fifth aspect of the present invention, the area dividing unit determines the area corresponding to the data value in the order from the maximum value of the two-dimensional or one-dimensional color space frequency to the minimum value of the frequency determined by the threshold value, and further determines the minimum value. Image data having a value less than or equal to the value is an area dividing unit which is a hue area dividing into one area.

According to a sixth aspect of the invention, the area dividing section has a peak detecting section for detecting a peak value of a frequency distribution of the two-dimensional or one-dimensional color space frequency, and an area for determining a hue area corresponding to the peak value. It is a divided part.

According to a seventh aspect of the present invention, the area dividing section is an area dividing section that divides the two-dimensional or one-dimensional color space so that the frequencies of the areas are equal.

According to the eighth aspect of the present invention, the image conversion unit calculates the density of the output image for each area obtained by dividing the color space area, or refers to the lookup table selected corresponding to the divided area. The density conversion method is used.

According to a ninth aspect of the present invention, the image conversion unit calculates the saturation of the output image for each area obtained by dividing the color space area, or uses the look-up table selected corresponding to the divided area. It is characterized in that it is a referenced saturation conversion method.

According to a tenth aspect of the present invention, the image conversion section comprises:
The hue conversion method is characterized in that the hue of the output image is calculated for each area obtained by dividing the color space area, or the hue conversion method is referred to by the lookup table selected corresponding to the divided area.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of an image enhancing apparatus according to the present invention will be described below with reference to the drawings.
In FIG. 1, an RGB image signal input from a scanner or the like is input to the coordinate conversion unit 1. The coordinate conversion unit 1 converts the RGB image signal into a signal (LHC) having three attributes of luminance (density), hue, and color vision of saturation. When converting RGB image signals into three attributes of color vision, FIG.
As shown in (a), the RGB image signal is once converted into C
L * a * b *, which is the IE color space (L * is lightness, a * b *
Is converted into hue and saturation) and then converted into LHC. From the RGB image signal, L
In order to directly obtain the 3-attribute signal without converting it into * a * b *, the formula of FIG. 2B is used. FIG. 2 (b)
, Max (R, G, B) is the set R, G, B,
It is a function that takes out the largest value from among. Also, min (R, G, B) is the set R, G, B,
It is a function that takes out the smallest value from among.
Further, mid (R, G, B) is a function for extracting the median value from the sets R, G, B.

In the coordinate conversion unit 1 shown in FIG. 1, an HLC (“HSV” when the three-attribute signal is directly formed from the RGB image signal) signal which is the three-attribute signal of color vision is obtained, and this HLC (or The HSV) signal is input to the histogram counting unit 2. The histogram counting unit 2 counts the histogram in two dimensions of H and C, or one dimension of H.

When the histogram is counted in two dimensions of H and C, the pixel data stored in the one-dimensional array variable (buf) is read pixel by pixel as shown in FIG. 3 (a). From the read data for one pixel, the hue value (buf (i) .a *) and the saturation value (buf
(I). b *) is obtained, and 1 is counted on a two-dimensional array variable (count) having the hue and saturation values as subscripts. Such processing is performed by using a one-dimensional array variable (bu
f) A two-dimensional histogram is obtained by applying it to all the pixel data stored above.

When the histogram is counted in only one dimension of H (hue), as shown in FIG. 3 (b),
The pixel data stored on the one-dimensional array variable (buf) is read pixel by pixel. The value (buf (i)) for one pixel read is obtained, and 1 is counted on a one-dimensional array variable (count) whose subscript is the value for this one pixel. Such processing is performed by using a one-dimensional array variable (bu
f) A one-dimensional histogram is obtained by applying it to all the pixel data stored above.

In FIG. 1, the histogram counting unit 2
The two-dimensional or one-dimensional histogram obtained in step 1 is input to the color area dividing unit 3. The color area dividing unit 3 divides the color area according to the input two-dimensional or one-dimensional histogram.

FIG. 7A conceptually shows division of the color area. In FIG. 7A, the color region is shown in two dimensions a * and b *.
The same applies to the two dimensions of S. In FIG. 7A, regions are formed for each dense color of pixels. It has been known by image recognition up to now that each area has similar colors and that it corresponds well to the pattern in the image. Note that FIG. 7B
Is a histogram corresponding to FIG.

Various methods can be considered for dividing the color area.
The simplest method is to select from the maximum pixel value in the histogram to the target area number K in order. At this time, the threshold of the histogram is determined, and the region is selected until the threshold is reached.

Another method for dividing the area is shown in FIG.
As shown in, in the two-dimensional histogram of hue and saturation, the five central, upper, lower, left and right variable values are checked, and the central value and the upper, lower, left and right 4 variables are checked. There is a method of comparing the values at the points and determining the peak value when the central value is the locally maximum pixel value and setting the peak value as the center of the area to obtain the area. In order to avoid noise, a threshold value may be set for the peak value and only peaks equal to or higher than the threshold value Pth may be determined as a region.

Further, as another area dividing method, FIG.
There is also something like that shown in In FIG. 5, first,
Of the two axes of hue and saturation that make up two dimensions, the length of the distribution of the histogram is measured, and the longer axis is focused on and divided into values at which the cumulative frequency is divided into two equal parts. In the divided areas, the axes having the longest distribution are searched again by comparing the two-dimensional axes, and the divided areas are again divided by the value at which the frequency accumulation is divided into two. This is repeated until the target number of regions Nth is reached or the remaining frequency accumulation becomes equal to or less than the threshold value Hth.

The partitioning method shown in FIG. 5 will be described in detail. As shown in FIG. 6, when the histogram distribution width wh0 on the H (hue) axis is longer than the axis widths of the other dimensions, as indicated by symbol A. The area is divided so that the cumulative distribution is divided into two equal parts. Thus, the axial length of the newly created region is w
h1l and wh1r. Next, the two axes of the lengths wh1l and wh1r generated by the division are compared with the axes of the other dimensions to find the longest distribution axis, and the cumulative distribution is calculated as 2
Divide into equal parts.

The area divided by each of the above division methods is input to the shortest distance search unit 4 as shown in FIG. The shortest distance search unit 4 measures the distance to which color region the pixel of the image signal is closest, and determines the closest region. Depending on the area to be determined, one look-up table (LUT) for conversion is selected, and the density,
Conversion for emphasizing saturation and hue is performed. Note that L
The N emphasis conversion may be performed using an arithmetic device without using the UT. By performing the conversion for emphasis in this way, the three-attribute signal L obtained from the coordinate conversion unit 1 is obtained.
HC becomes a three-attribute signal L'H'C 'in which density, saturation, and hue are emphasized. The 3-attribute signal L'H'C 'is input to the coordinate conversion unit 11, converted back to an RGB image signal, and input to a color correction circuit or the like.

Next, conversion to optimum density and saturation will be described. As shown in FIG. 8A, if the distribution of the histogram of each axis is biased, the density and saturation are biased, resulting in a flat and unsharp image. Therefore, if the distribution of such a biased histogram can be averaged as shown in FIG. 7B, the image is contrasted and emphasized.

To average the biased histogram distribution, a method of linearly stretching the histogram is used. To linearly stretch the histogram, D = F {255 / (DATAmax-DATAmin) * (DATA-DATA Amin)}. . . The formula of (1) is used. In the above formula (1), DATAmax
Represents the maximum value of the density or saturation in the original, DATAmin represents the minimum value of the density or saturation in the original, and DATA represents the histogram data. Further, F is an arbitrary constant.
By substituting all the histogram data into the equation (1) and converting it, the histogram is linearly stretched and the distribution is averaged. Then, the image is contrasted and enhanced.

As a method of averaging the biased histogram distribution, there is also a method of uniformly distributing the histogram data over the entire dynamic range to improve the contrast. For example, let D be the first number of quantization steps in each of the two-dimensional axes, D ′ be the smaller number of requantization steps, and M be the average value obtained by averaging the histogram data. Then, the variable for operating the array is set to i, and i is 0.
The counts, which are histogram data, for the cumulative sum SUM (one-dimensional array variable) in order from D to D
Continue adding (i). In addition, the cumulative SUM is the average value M
It is repeated until. When the addition to the cumulative SUM is completed, the subscript of the cumulative SUM is incremented by 1, and the count (i) is continuously added until the average value M is reached. When this reaches the target quantization number D ′, the requantized histogram of each step is flattened to almost the same number.

To enhance the saturation, the above histogram may be flattened or linearized. However, in a general copying machine or printer, the saturation is lowered with respect to the original, so C = aC + b.
(A and b are constants) are used to enhance the saturation.

Hue emphasis is applied to memory colors such as flesh color, green of plants and blue of sky, and is shifted in a preferred color direction for each color region. For example, the skin color is shifted to pink (magenta) from the original color tone.

According to the image emphasizing apparatus having the above-mentioned configuration, the conversion section 1 for converting the RGB image data into the three-attribute signal, the histogram counting section 2 for obtaining a two-dimensional or one-dimensional histogram from the three-attribute signal, Since the area dividing unit 3 that divides the histogram into a plurality of regions and the image converting unit that determines the optimum density, saturation, and hue based on the regions divided by the region dividing unit 3 are provided, it is optimum for each pattern region. It is possible to obtain a high contrast. Furthermore, since the density, saturation, and hue can be optimized separately, different optimum emphasis can be applied to each picture. For example, the saturation may be emphasized less in the flesh-colored area, and conversely, the saturation may be more emphasized in order to increase the vividness in the green area.

[0041]

According to the present invention, a coordinate conversion unit that obtains at least two-dimensional or more data in which the color space of image data represents a color, and a plurality of regions are created based on the appearance frequency distribution of the two-dimensional or more data. Since the area dividing unit for classifying and the image converting unit for determining the conversion amount of the density, hue, and saturation of the output image are provided, it is possible to obtain the optimum contrast for each pattern region. Furthermore, since the density, saturation, and hue can be optimized separately, it is possible to apply different optimum emphasis for each picture.

[Brief description of drawings]

FIG. 1 is a conceptual diagram showing an embodiment of an image enhancing apparatus according to the present invention.

FIG. 2 is a diagram showing an equation for obtaining a 3-attribute signal from an RGB-based image signal in the image enhancing apparatus.

FIG. 3 is a diagram showing a logic for forming an appearance frequency distribution from an image signal in the image enhancing apparatus.

FIG. 4 is a diagram showing an example of area division in the image enhancement apparatus.

FIG. 5 is a diagram showing another example of area division in the image enhancement apparatus.

FIG. 6 is a diagram conceptually showing the area division shown in FIG.

FIG. 7 is a diagram conceptually showing division of a color area of image data.

FIG. 8 is a diagram showing a histogram distribution of image data,
(A) is a biased distribution, and (b) is an averaged distribution.

FIG. 9 is a diagram conceptually showing an example of a conventional image forming apparatus.

FIG. 10 is a diagram showing a color gamut area of a display and a printer.

[Explanation of symbols]

 1 conversion unit 2 histogram count unit

Claims (10)

[Claims] 1. A coordinate conversion unit for obtaining at least two-dimensional or more data representing a color from a color space of image data, and a region dividing unit for classifying into a plurality of regions based on the appearance frequency distribution of the two-dimensional or more data. And an image conversion unit that determines a conversion amount of density, hue, and saturation of an output image. 2. The image conversion unit converts a data value in an RGB color space into an L * a * b * color space, and further, a
Two-dimensional color space of hue and saturation determined from the * b * plane,
Alternatively, the image enhancing apparatus according to claim 1, wherein the image enhancing apparatus is a one-dimensional color space of hue. 3. The image conversion unit is a two-dimensional color space of hue and saturation directly calculated by calculation from a data value of the RGB color space, or a one-dimensional hue space of hue. Item 1. The image enhancing device according to item 1. 4. The area dividing unit is a means for dividing the two-dimensional color data frequency distribution indicating the tint of the input image or the one-dimensional hue frequency distribution into an integer number of color space areas. The image enhancing device according to claim 1. 5. The area dividing unit determines an area corresponding to a data value in an order from the maximum value of the two-dimensional or one-dimensional color space frequency to the minimum value of the frequency determined by a threshold value, and further, the image data having the minimum value or less is determined. The image enhancement apparatus according to claim 1, wherein is an area dividing unit that is a hue area division into one area. 6. The area dividing section has a peak detecting section for detecting a peak value of a frequency distribution of a two-dimensional or one-dimensional color space frequency, and is a region dividing section for determining a hue area corresponding to the peak value. The image enhancing device according to claim 1, wherein 7. The image enhancing apparatus according to claim 1, wherein the region dividing unit is a region dividing unit that divides the two-dimensional or one-dimensional color space so that the frequencies of the regions are equal. . 8. The image conversion unit calculates the density of an output image for each area obtained by dividing a color space area, or the density conversion method referred to by a lookup table selected corresponding to the divided area. The image enhancement device according to claim 1, wherein 9. The image conversion unit calculates the saturation of an output image for each area obtained by dividing a color space area, or the saturation referred to in a lookup table selected corresponding to the divided area. The image enhancement apparatus according to claim 1, wherein the image enhancement apparatus is a conversion method. 10. The image conversion unit calculates a hue of an output image for each area obtained by dividing a color space area, or a hue conversion method referred to by a look-up table selected corresponding to the divided area. 2. The method according to claim 1, wherein
The image enhancement device described.