JP3879720B2 - Image processing apparatus and image processing method - Google Patents

Description

  The present invention relates to a full color printing-related device such as a printer, a video printer, or a scanner, an image generation device that generates a computer graphics image, or a color display device, and the frequency characteristics of color image data are matched to the device or use conditions. The present invention relates to an image processing apparatus and an image processing method for conversion.

  Image data input to an image display device or the like is generally composed of R, G, and B color data. Such color data is generated by an image generation device such as a camera. However, since it is affected by various noises in the transmission process, the color data input to the image display device may not match the original color data. . For this reason, processing such as removal of noise components is performed on the image display apparatus side so that the original image is reproduced.

The following patent document is mentioned as an example of the image processing apparatus which smoothes an image. The image processing apparatus described in this patent document discriminates a dark part and a bright part of an image from input image data, and performs smoothing processing only on a bright part adjacent to the dark part. When displayed, the line can be prevented from appearing thin.
JP 2002-41025 A

In a color image, it is necessary to perform frequency characteristic conversion processing such as noise component removal for each color component. For example, if a noise component is included in color data representing human skin color, it appears as wrinkles. In such a case, it is desirable to remove a noise component included only in the skin color region. Since conventional noise removal processing is uniformly performed for all color data of R, G, and B, the same frequency component as the noise component is removed even for a color where noise is not noticeable, resulting in blurring of the image and the like. appear. Also, if the original image has little change in the red area (that is, only low frequency components are present), and the blue area is severely changed (that is, high frequency components are present), the red area in the transmission process When a noise component is included in the image, it is visually recognized and the image quality is deteriorated. In this case, it is desirable to remove noise components only in the red region and not perform noise removal in the blue region.

  In addition, human visual characteristics include colors that are very sensitive to changes and colors that are not very sensitive. A viewer visually recognizes a color image as a combination of color (light) stimuli displayed based on R, G, and B color data, but the amount of noise perceived by the viewer varies depending on the hue, and is human. There are hues that are easy to feel noise and hues that are difficult to feel due to the visual characteristics. Therefore, it is not appropriate to perform similar noise removal for all hues.

  The present invention has been made in view of the above problems, and an object thereof is to provide an image processing apparatus and an image processing method capable of independently converting the frequency characteristics of a specific hue component in a color image.

An image processing apparatus according to the present invention is an image processing apparatus that converts frequency characteristics of first color data representing a color image and outputs second color data corresponding to the first color data.
Hue area data calculating means for calculating first hue area data effective for a plurality of specific hue components in the color image represented by the first color data, using the first color data;
Frequency characteristic conversion means for outputting the second hue area data by independently converting the frequency characteristics of the first hue area data for each of the hue components;
Synthesis means for calculating the second color data using the second hue area data.

An image processing method according to the present invention converts the frequency characteristics of first color data representing a color image and outputs second color data corresponding to the first color data,
Using the first color data, calculating first hue area data effective for a plurality of specific hue components in the color image represented by the first color data,
The second hue region data is output by converting the frequency characteristics of the first hue region data independently for each of the hue components,
The second color data is calculated using the second hue area data.

  According to the image processing device and the image processing method of the present invention, the first hue area data effective for a plurality of specific hue components in the color image represented by the first color data is calculated, and the first Since the second color data is calculated using the second hue area data obtained by converting the frequency characteristics of one hue area data for each hue component, the frequency characteristics of the first color data are Each hue component can be controlled independently.

  Hereinafter, an image processing apparatus according to the present invention will be specifically described with reference to the drawings.

  FIG. 1 is a block diagram showing an embodiment of an image processing apparatus according to the present invention. As shown in FIG. 1, the image processing apparatus according to the present embodiment includes an αβ calculating means 1, a chromatic color component data calculating means 2, a hue area data calculating means 3, frequency characteristic converting means 4a to 4f, and a synthesizing means 5. Is done.

  First color data Ri, Gi, Bi representing a color image is input to αβ calculating means 1 and chromatic color component data calculating means 2. The αβ calculating means 1 selects and outputs the maximum value β and the minimum value α of the first color data Ri, Gi, Bi. The output maximum value β and minimum value α are input to the chromatic color component data calculation means 2. The minimum value α is also input to the synthesis means 5. Here, the minimum value α is data representing the size of the achromatic color (gray) component included in the first color data Ri, Gi, Bi.

  The chromatic color component data calculation means 2 is represented by first color data based on the first color data Ri, Gi, Bi and the maximum value β and the minimum value α output from the αβ calculator means 1. The chromatic color component data r, g, b, y, m, c representing the size of each color component of red, green, blue, yellow, magenta, cyan of the color (chromatic color) obtained by removing the achromatic color component from the color. calculate. These chromatic color data are obtained by subtraction processing of r = Ri−α, g = Gi−α, b = Bi−α, y = β−Bi, m = β−Gi, and c = β−Ri.

  5A to 5F schematically show the sizes of chromatic color component data y, m, c, r, g, and b in six hues of red, yellow, green, cyan, blue, and magenta. It is a figure. As shown in FIG. 5, each of the chromatic color component data is maximum in three hues among the six hues and zero in the remaining three hues. For example, the chromatic color component data c has a maximum value in the hues of green, cyan, and blue, the size decreases from the hue of green to yellow, and the hue of blue to magenta, and in the hues of red, yellow, and magenta. Is zero.

  The chromatic color component data obtained as described above has a property that at least one of r, g, and b and at least one of y, m, and c are zero. For example, when the maximum value β is Ri and the minimum value α is Gi (β = Ri, α = Gi), g = 0 and c = 0 are obtained by the above subtraction process, and the maximum value β is Ri and the minimum value α. Is Bi (β = Ri, α = Bi), b = 0 and c = 0. That is, depending on the combination of Ri, Gi, and Bi that is the maximum and minimum, at least one of r, g, and b, and any one of y, m, and c in total will be zero.

  Six chromatic color component data r, g, b, y, m, and c output from the chromatic color component data calculation means 2 are sent to the hue area data calculation means 3. FIG. 2 is a block diagram illustrating an example of an internal configuration of the hue area data calculation unit 3. The hue area data calculation unit 3 includes a plurality of minimum value selection units 6a to 6f that select and output the smaller value of the two input chromatic color component data.

The minimum value selection means 6a selects the smaller value of the chromatic color component data r and b and outputs it as the hue area data h1m. Similarly, the minimum value selecting unit 6b selects the smaller one of the chromatic color component data r and g and outputs it as the hue area data h1y, and the minimum value selecting unit 6c selects the smaller one of the chromatic color component data g and b. The minimum value selection means 6d selects the smaller one of the chromatic color component data y and c and outputs it as the hue area data h1g, and the minimum value selection means 6e is the chromatic color component data. The smaller one of y and m is selected and output as hue area data h1r, and the minimum value selection means 6f selects the smaller of chromatic color component data m and c and outputs it as hue area data h1b.

The calculation of the hue area data h1r, h1g, h1b, h1c, h1m, h1y can be expressed by the following equation.
h1r = min (y, m) (1-1) h1y = min (r, g) (1-4)
h1g = min (y, c)… (1-2) h1m = min (r, b)… (1-5)
h1b = min (m, c)… (1-3) h1c = min (m, c)… (1-6)
However, min (A, B) represents the smaller value of A and B.

  6A to 6F schematically show the relationship between the hue area data h1r, h1y, h1g, h1c, h1b, and h1m and the six hues of red, yellow, green, cyan, blue, and magenta. It is a figure. As shown in FIG. 6, the hue area data h1r, h1g, h1b, h1c, h1m, and h1y are the maximum in the hues of red, green, blue, cyan, magenta, and yellow, respectively, and the sizes in the other hues are large. 0. That is, the hue area data h1r, h1g, h1b, h1c, h1m, and h1y are data effective for hue components of red, green, blue, cyan, magenta, and yellow in the color image represented by the first color data. I can say that.

For example, when W is a constant and the first color data is Ri = W, Gi = 0, Bi = 0, the first color data represents the color of red hue. At this time, the chromatic color component data is r = W, g = b = 0, y = m = W, and c = 0. Therefore, h1r = min (y, m) = W, and the other five hue area data h1g, h1b, h1y, h1m, and h1c are all 0. That is, for the red hue, only h1r = min (y, m) is effective hue area data. Similarly, h1g = min (c, y) for green, h1b = min (m, c) for blue, h1c = min (g, b) for cyan, and h1m = min (b, r) for magenta. For yellow, h1y = min (r, g) is effective hue area data.

  The hue area data h1r, h1g, h1b, h1y, h1m, and h1c calculated by the hue area data calculating unit 3 are input to the frequency characteristic converting units 4a, 4b, 4c, 4d, 4e, and 4f, respectively. The frequency characteristic conversion means 4a to 4f convert the spatial frequency characteristics and / or temporal frequency characteristics of the input hue area data, and the converted hue area data fh1r, fh1g, fh1b, fh1c, fh1m, fh1y. Is output.

  The frequency conversion characteristics of the frequency characteristic conversion means 4a, 4b, 4c, 4d, 4e, 4f are set according to the input hue area data h1r, h1g, h1b, h1y, h1m, h1c, respectively. For example, when noise in a color component related to a specific hue is removed, the frequency characteristic conversion means to which the hue area data corresponding to the specific hue is input removes (attenuates) the frequency component in which the noise component exists. Configured as a filter.

  Further, when reproducing a smooth image, it is constituted by a low-pass filter that blocks or attenuates high-frequency components of input hue area data and transmits low-frequency components. Furthermore, when emphasizing the contour portion of the image, the frequency characteristic converting means is configured by a filter that emphasizes the contour portion of the input hue area data, that is, the high frequency component.

  Hereinafter, the case where the noise component of the hue area data is removed by the frequency characteristic conversion means will be described. In this case, the conversion characteristic of the frequency characteristic conversion means is determined according to the characteristic of the noise component included in the hue area data. Noise includes white noise composed of components in a wide frequency range from low frequency to high frequency, and noise having a strong component at a specific frequency caused by the influence of a carrier wave during transmission.

  In the case of noise having a strong component at a specific frequency, the frequency characteristic conversion means is configured as a filter that removes or attenuates the frequency component. When the noise component included in the image data is white noise, the frequency characteristic conversion unit is configured as a filter that removes or attenuates a visually noticeable frequency component among the noise components. Strictly speaking, such visually conspicuous frequency components are specified by the pixel interval or viewing distance of the display device, but in a general image display device, many noise components close to the pixel frequency of image data are seen. Here, the pixel frequency refers to a frequency when the image data of each successive pixel changes to, for example, 0, 1, 0, 1, 0..., And corresponds to ½ of the clock frequency of the image data. In a pixel interval of a general image display device currently used, noise components in which data of adjacent pixels frequently change unrelated to each other are visually recognized, so it is desirable to remove them.

  As described above, when the noise of the image data is removed by the frequency characteristic conversion means 4a to 4f, the filter characteristic is determined according to the frequency band of the noise component to be removed. Hereinafter, the case where the noise component in the high frequency region close to the pixel frequency of the image data is removed by the frequency characteristic conversion means 4a to 4b will be described. In this case, the frequency characteristic conversion means 4a to 4f can be configured by a low-pass filter that blocks or attenuates high-frequency components in image data and transmits low-frequency components. Specifically, a low-pass filter that cuts off or attenuates a frequency component at about 1 / 4.5 or more of the pixel frequency, that is, 1/9 or more of the clock frequency of the pixel data is used. As a simple configuration, a low-pass filter that calculates a simple average value of a plurality of consecutive pixels, which will be described below, can be considered. In this case, the filter characteristic is determined by the number of pixels used for calculating the simple average value.

FIG. 3 is a block diagram showing the internal configuration of each of the frequency characteristic conversion means 4a to 4f. As shown in FIG. 3, the frequency characteristic conversion units 4 a to 4 f include a data shift unit 8 including a plurality of data storage units 7 a to 7 h and a weighted addition unit 9. The hue area data h1r, h1g, h1b, h1y, h1m, and h1c input to each of the frequency characteristic conversion units 4a to 4f are sent to the data storage unit 7a. The data storage units 7a to 7h are cascade-connected to each other, and each data storage unit 7a to 7h shifts the input data all at once to the subsequent stage every time the hue area data is input, and Output.
When the first data is input to the data shift unit 8, the same data is input to the data storage units 7a to 7h all at once. When the last data is input, the same data is continuously input to the data storage units 7a to 7h.

  The weighted addition means 9 performs weighted addition on the data output from the data storage units 7a to 7h, and outputs the result of the weighted addition as converted hue area data fh1r, fh1g, fh1b, fh1y, fh1m, and fh1c. When the weighting and adding means 9 performs weighting evenly, that is, if the weighting coefficients are set to the same value, a simple average value is calculated. In this case, the post-conversion hue area data fh1r calculated by the frequency characteristic conversion unit 4a is expressed by the following equation.

fh1r = f (hir [n + 4], hir [n + 3], hir [n + 2], hir [n + 1], hir [n], hir [n-1]
, hir [n-2], hir [n-3], hir [n-4]) ... formula (2)

In the above formula (2), h1r [n] represents the nth input hue area data, and the function f is the data hir [n + 4]... Hir [n-4] output from the data storage units 7a to 7h. Represents weighted addition.
The other converted hue area data fh1g, fh1b, fh1c, fh1m, and fh1y are also expressed in the same manner as the above equation (2).

  The converted hue area data fh1r, fh1g, fh1b, fh1c, fh1m, and fh1y are input to the synthesizing unit 5 together with the minimum value α. The synthesizing unit 5 calculates the second color data Ro, Go, Bo by synthesizing the converted hue data and the minimum value α.

  FIG. 4 is a block diagram showing the internal configuration of the synthesizing means 5. As shown in FIG. 4, the synthesizing means 5 is composed of adding means 10a to 10f. The adding means 10a adds the converted hue area data fh1r, fh1y, fh1m, and calculates the converted chromatic color component data r1. Further, the adding means 10b adds the converted hue area data fh1g, fh1y, fh1c, and calculates the converted chromatic color component data g1. The adding means 10c adds the converted hue area data fh1b, fh1c, and fh1m to calculate the converted chromatic color component data b1. The adding means 10d adds the converted chromatic color component data r1 and the minimum value α to calculate the second color data Ro. Further, the adding means 10e adds the converted chromatic color component data g1 and the minimum value α to calculate the second color data Go, and the adding means 10c calculates the converted chromatic color component data b1 and the minimum value α. Addition is performed to calculate the second color data Bo.

The above calculation in the synthesizing means 5 is expressed by the following equation.
Ro = fh1r + fh1m + fh1y + α ... Formula (3-1)
Go = fh1g + fh1y + fh1c + α… Formula (3-2)
Bo = fh1b + fh1c + fh1m + α… Formula (3-3)

  FIGS. 7A to 7C schematically show six hues and post-conversion chromatic color component data r1, g1, and b1. For simplification, the chromatic color component data r1, g1, and b1 after conversion shown in FIGS. 7A to 7C do not represent the influence of frequency characteristic conversion in the frequency characteristic conversion units 4a to 4f. That is, fh1r = h1r, fh1g = h1g, fh1b = h1b, fh1c = h1c, fh1m = h1m, and fh1y = h1y. In this case, the converted chromatic color component data is the same as the chromatic color component data r, g, b The second color data Ro, Go, Bo obtained by adding the minimum value α is the same as the first color data.

  As described above, the image processing apparatus according to the present invention independently outputs the hue area data h1r, h1g, h1b, h1y, h1m, and h1c effective for the hue components of red, green, blue, yellow, magenta, and cyan, respectively. Since the second color data is calculated by synthesizing the converted hue area data fh1r, fh1g, fh1b, fh1y, fh1m, and fh1c obtained by conversion, the hue components of red, green, blue, yellow, cyan, and magenta The frequency characteristics can be converted every time.

  Hereinafter, the operation of the frequency characteristic conversion means 4a to 4f will be described. The first color data Ri, Gi, Bi are affected by various noises in the transmission process. The first color data Ri to be input to the image display device is assumed when the original color components of the first color data are Rs, Gs, Bs, and the noise components are Rn, Gn, Bn. , Gi, Bi can be expressed as Ri = Rs + Rn, Gi = Gs + Gn, Bi = Bs + Bn. That is, the first color data Ri, Gi, Bi input to the image display device is represented by the sum of the original color data components Rs, Gs, Bs and the noise components Rn, Gn, Bn. Will be.

  FIG. 8 is a diagram illustrating an example of the size of the first color data Ri, Gi, Bi input to the image processing apparatus. In FIG. 8, the horizontal axis represents the pixel position, and the vertical axis represents the size of each of the first color data Ri, Gi, Bi at each pixel position. The original color data of the first color data Ri, Gi, Bi shown in FIG. 8 is Rs = 160, Gs = 32, Bs = 32 (including a gray component) at all pixel positions, and represents a uniform red color. . However, due to the influence of the noise components Rn, Gn, Bn, the first color data Ri, Gi, Bi input to the image processing apparatus are not uniform as shown in FIG.

  Here, the frequency characteristic conversion unit 4a performs noise removal processing on the hue region data h1r, and the frequency characteristic conversion units 4b to 4f perform hue region data h1g, h1b, h1y, input without performing noise removal processing. Assume that h1m and h1c are output as they are. That is, a simple average value is calculated by the weighting addition means 9 of the frequency characteristic conversion means 4a, and weighting addition is not performed in the weighting addition means 9 of the other frequency characteristic conversion means 4b to 4f. In this case, the effect of frequency characteristic conversion appears only in the converted hue region data fh1r effective for the red hue component.

  FIG. 9 is a diagram showing second color data Ro, Go, Bo calculated by performing the above-described processing on the first color data Ri, Gi, Bi shown in FIG. As shown in FIG. 9, it can be seen that the influence of the noise component is small for Ro among the second color data Ro, Go, and Bo. On the other hand, the influence of the noise component does not change for Go and Bo. This is because the characteristics of the frequency characteristic conversion means 4a to 4f are determined so that the effect of the frequency characteristic conversion appears only on the hue area data h1r effective for the red hue component. At this time, in the frequency characteristic conversion means 4a, noise removal processing is performed only on the hue area data h1r effective for the red hue component, so that the red hue is not affected without affecting the hues of adjacent yellow and magenta. Noise removal processing can be performed only for the components.

  Similarly, in the frequency characteristic conversion means 4b, the weighting coefficient of the weight addition means 9 is set so that noise removal processing is performed on the hue area data h1g effective for the green hue component, and the other frequency characteristic conversion means 4a, If the input hue area data h1r, h1b, h1y, h1m, and h1c are output as they are without performing noise removal processing in 4c to 4f, the influence of the noise component is exerted only on the second color data Go. Can be small. At this time, since the noise removal processing is performed only on the hue area data h1g effective for the green hue component in the frequency characteristic conversion means 4b, the green color is not affected without affecting the adjacent yellow and cyan hue components. Noise removal processing can be performed only on the hue component.

  As described above, the image processing apparatus according to the present invention uses the hue area data h1r, h1y, h1g, h1c, h1b, effective for the hue components of red, yellow, green, cyan, blue, and magenta from the first color data. Since h1m is generated and frequency characteristic conversion processing is performed for each hue area data, the frequency characteristics of the first color data can be controlled independently for each of the hue components of red, yellow, green, cyan, blue, and magenta. it can. As a result, when the influence of the noise component in a specific hue is conspicuous, or when it is desired to remove only the noise component in a specific hue, only the noise component in the specific hue is affected without affecting the frequency characteristics of other hues. It can be removed. For example, since a noise component included in human skin color is visually recognized as a wrinkle, noise removal may be performed on red and / or yellow hue area data in order to remove the noise component present in the skin color.

  Also, if the original image shows little change in the red hue (less high-frequency component) and a great change in the blue hue (more high-frequency component), the red hue will When a noise component is mixed, the noise component is very noticeable. In this case, noise processing can be effectively performed by removing the noise component only for the hue area data h1r effective for the red hue.

  In addition, human visual characteristics include colors that are very sensitive to changes and colors that are not very sensitive. Image data generally consists of red, green, and blue color data, and a noise component is also added to each of the three color data. At this time, even if the magnitude of the noise component added to each color data is the same, the magnitude of the noise visually recognized by the viewer changes for each hue depending on the human visual characteristics, There will be a hue that is difficult to feel. According to the image processing apparatus in the present embodiment, a visually superior image can be obtained by determining the noise removal characteristics for each hue in consideration of human visual characteristics.

  In addition, in the weighting addition means 9 of the frequency characteristic conversion means 4a to 4f of the image processing apparatus in the present embodiment, the conversion characteristics can be flexibly changed by appropriately changing the weighting coefficient. In the image processing apparatus according to the present embodiment, the frequency characteristic conversion unit is provided for all hue area data. However, the frequency characteristic conversion unit is provided only for hue area data effective for the hue of interest. Therefore, the apparatus can be downsized.

  In the image processing apparatus according to the present embodiment, the frequency characteristic conversion units 4a to 4f are configured to convert spatial frequency characteristics by calculation between hue area data of adjacent pixels. It is good also as a structure which performs conversion of a frequency characteristic. In other words, storage means for storing the hue area data for several screens is provided, and the temporal frequency characteristics are also converted independently for each hue by performing calculations between the hue area data for several screens before and after the same pixel. Is possible.

  FIG. 10 is a block diagram showing another embodiment of the image processing apparatus according to the present invention. As shown in FIG. 10, the image processing apparatus according to the present embodiment performs frequency characteristic conversion means 4g that performs processing on the minimum value α representing the magnitude of the achromatic color component of the first color data Ri, Gi, Bi. Is further provided. Other configurations are the same as those of the image processing apparatus according to the first embodiment shown in FIG.

The frequency characteristic conversion means 4g performs a spatial or temporal frequency characteristic conversion process on the minimum value α, and outputs a converted minimum value fα. The frequency characteristic conversion unit 4g can be configured in the same manner as the other frequency characteristic conversion units 4a to 4f shown in FIG. 3, and according to the set value of the weighting coefficient, a noise removal unit, a low-pass filter, a contour enhancement unit, a high frequency Acts as a highlighting means.
As described above, the minimum value α is data representing the size of the achromatic component in the first color data Ri, Gi, Bi. Therefore, the frequency characteristic conversion means 4g performs frequency characteristic conversion on the achromatic color component.

  The image processing apparatus according to the present embodiment changes the frequency conversion characteristics of the first color data for each hue area data by the frequency characteristic conversion means 4a to 4f and converts the frequency characteristics of the achromatic color component independently. It can be carried out.

  FIG. 11 is a block diagram showing another embodiment of the image processing apparatus according to the present invention. In the image processing apparatus shown in FIG. 11, in addition to the converted hue area data fh1r, fh1g, fh1b, fh1c, fh1m, fh1y, and the minimum value α, the hue area data h1r, h1g, h1b, h1c, h1m, h1y are combined. Input to means 5b. The synthesizing unit 5b generates second color data Ro, Go, Bo by weighted addition of the converted hue data and the hue area data. Other configurations are the same as those of the image processing apparatus according to the first embodiment shown in FIG.

  FIG. 12 is a block diagram showing the internal configuration of the synthesizing means 5b. The weighted addition means 9a to 9f perform weighted addition of the hue area data and the converted hue area data. That is, the weighting addition means 9a weights and adds the hue area data h1r and the converted hue area data fh1r, and the weighting addition means 9b weights and adds the hue area data h1y and the converted hue area data fh1y, and the weighting addition means 9c. Is a weighted addition of the hue area data h1g and the converted hue area data fh1g, the weighting addition means 9d is a weighted addition of the hue area data h1c and the converted hue area data fh1c, and the weighted addition means 9e is the hue area data h1b. The converted hue area data fh1b is weighted and added, and the weighting adding means 9f weights and adds the hue area data h1m and the converted hue area data fh1m.

  The adder 10a adds the outputs from the weighted adders 9a, 9b and 9f, and calculates the converted chromatic color component data r1. Further, the adding means 10b adds the outputs from the weighted adding means 9b, 9c and 9d, and calculates the converted chromatic color component data g1. Further, the adding means 10c adds the outputs from the weighted adding means 9d, 9e and 9f, and calculates the converted chromatic color component data b1.

  The adding unit 10d adds the converted chromatic color component data r1 and the minimum value α to calculate the second color data Ro. Further, the adding means 10e adds the converted chromatic color component data g1 and the minimum value α to calculate the second color data Go. Further, the adding means 10c adds the converted chromatic color component data b1 and the minimum value α to calculate the second color data Bo.

The above calculation in the synthesizing means 5 is expressed by the following equation.
Ro = (kr × fh1r + (1-kr) × h1r) + (km × fh1m + (1-km) × h1m)
+ (ky × fh1y + (1-ky) × h1y) + α ... Formula (4-1)
Go = (kg × fh1g + (1-kg) × h1g) + (ky × fh1y + (1-ky) × h1y)
+ (kc × fh1c + (1-kc) × h1c) + α… Formula (4-2)
Bo = (kb × fh1b + (1-kb) × h1b) + (kc × fh1c + (1-kc) × h1c)
+ (km × fh1m + (1-km) × h1m) + α… Formula (4-3)
Here, kr, kg, kb, kc, km, ky are weighting coefficients, which are values of 0 or more and 1 or less.

Hereinafter, the effect of weighted addition of the hue area data and the converted hue area data in the synthesizing unit 5b will be described.
FIG. 13 is a diagram illustrating an example of the first color data Ri, Gi, Bi input to the image processing apparatus according to the present embodiment illustrated in FIG. In FIG. 13, the horizontal axis represents the pixel position where each data exists, and the vertical axis represents the size of each data. The original color data is Rs = 32, Gs = 32, and Bs = 32 at pixel positions 0 to 42 and represents uniform gray, and after pixel position 43, Rs = 160, Gs = 32, and Bs = 32. Yes, representing a uniform red color with a gray component. However, since the first color data Ri, Gi, Bi are affected by noise components, they are not uniform as shown in FIG.

  Here, the frequency characteristic conversion unit 4a performs noise removal processing on the hue region data h1r, and the frequency characteristic conversion units 4b to 4f perform hue region data h1g, h1b, h1y, input without performing noise removal processing. Assume that h1m and h1c are output as they are. That is, a simple average value is calculated by the weighting addition means 9 of the frequency characteristic conversion means 4a, and weighting addition is not performed in the weighting addition means 9 of the other frequency characteristic conversion means 4b to 4f. In this case, the effect of frequency characteristic conversion appears only for the hue area data effective for red.

  FIG. 14 is a diagram showing second color data Ro, Go, Bo calculated by performing the above-described processing on the first color data Ri, Gi, Bi shown in FIG. 13, and is shown in FIG. Each of the cases where the weighting coefficient kr in the weighting addition means 9a of the synthesizing means 5b is set to 0, 0.5, 1 is shown. As shown in FIG. 14, it can be seen that the slope of the changed portion of the second color data Ro is changed by changing the weighting coefficient kr. That is, in the case of kr = 1, the change at the boundary portion becomes gradual, and as a result, “color leakage” in which the portion before the pixel position 42 that was originally gray is colored red occurs. Further, when kr = 0.5, “color leakage” before the pixel position 42 is reduced. When kr = 0, it is the same as when frequency characteristic conversion is not performed.

  As described above, the smaller the value of the weighting coefficient kr, that is, the smaller the influence of the converted hue area data fh1r compared to the hue area data h1r, the smaller the effect of the frequency characteristic conversion. Therefore, it is possible to adjust the magnitude of the effect of the frequency characteristic conversion for each hue by making the conversion characteristics in the frequency characteristic converting means 4a to 4f the same and changing the weighting coefficient in the synthesizing means 5b for each hue. .

  As described above, according to the image processing apparatus in the present embodiment, by adjusting the value of the weighting coefficient in the synthesizing unit 5b, the “color” to other hues can be changed without changing the conversion characteristic in the frequency characteristic converting unit. The degree of “leakage” can be adjusted. In addition, by changing the weighting coefficient in the synthesizing unit 5b for each hue, the magnitude of the effect of the frequency characteristic conversion can be adjusted for each hue.

  FIG. 15 is a block diagram showing another embodiment of the image processing apparatus according to the present invention. As shown in FIG. 15, in the image processing apparatus according to the present embodiment, a data selection unit 11 is provided after the frequency characteristic conversion units 4a to 4f. The αβ calculating means 1a selects and outputs the maximum value β and the minimum value α of the first color data Ri, Gi, Bi, and the color represented by the first color data Ri, Gi, Bi. An identification code S1 representing information relating to the hue is output.

  FIG. 16 is a diagram showing the relationship between the identification code S1 and the first color data Ri, Gi, Bi maximum value β, minimum value α, and hue data of zero. As shown in FIG. 16, identification codes S1 of 0 to 12 are provided according to combinations of the maximum value α and the minimum value β of Ri, Gi, Bi. Here, the hue information of the first color data is specified by the relationship between the maximum value α and the minimum value β of Ri, Gi, B. For example, when the maximum value α = Ri and the minimum value β = Gi = Bi, the first color data represents the color of the red hue.

FIG. 17 is a diagram illustrating the relationship between the value of the identification code S1, the size of the first color data Ri, Gi, Bi, and the hue. As shown in FIG. 16, when the first color data Ri, Gi, Bi represents a hue color from red to yellow, 1 is output as the identification code S1. Similarly, S1 = 3 for yellow to green, S1 = 2 for green to cyan, S1 = 4 for cyan to blue, S1 = 5 for blue to magenta, S1 for magenta to red. = 0 is output as an identification code. Further, when the color data Ri, Gi, Bi of the first color represents a red hue, 6 is output as the identification code S1. Similarly, S1 = 11 for yellow, S1 = 7 for green, S1 = 9 for cyan, S1 = 8 for blue, and S1 = 10 for magenta. .
When Ri = Gi = Bi, the first color data represents an achromatic color, that is, gray. In this case, 12 is output as the identification code S1.

  The identification code S1 is output to the data selection means 11. The data selection unit 11 selects the converted hue area data fh1r, fh1g, fh1b, fh1c, fh1m, and fh1y output from the frequency characteristic conversion units 4a to 4f based on the value of the identification code S1, and the selected conversion The post-hue area data sfh1r, sfh1g, sfh1b, sfh1c, sfh1m, and sfh1y are output.

  FIG. 18 is a diagram illustrating the relationship between the identification code S1 and the converted hue area data fh1r, fh1g, fh1b, fh1c, fh1m, and fh1y selected based on the identification code S1. For example, when S1 = 1, the first color data Ri, Gi, Bi represent hue colors from red to yellow, so only the hue area data h1r effective for red and the hue area data h1y effective for yellow are included. It has a non-zero value, and the other hue area data h1g, h1b, h1c, h1m are 0.

  However, the converted hue area data becomes non-zero in the area that should originally be zero by the noise removal processing in the frequency characteristic conversion means 4a to 4f (simple average calculation in the weighted addition means 9). In such a region that should originally be zero, color leakage occurs due to non-zero hue region data after conversion. In order to prevent this problem, the data selection unit 11 selects the converted hue area data based on the value of the identification code S1 so that the converted hue area data that should be zero is zero. That is, when the identification code S1 = 1, the hue area data h1r effective for red and the hue area data h1y effective for yellow are selected and output as sfh1r = fh1r and sfh1y = fh1y. It outputs as sfh1g = sfh1b = sfh1c = sfh1m = 0 so that all become zero. Similarly, when the identification code S1 = 6, the converted hue area data h1r effective for red is selected to be sfh1r = fh1r, and is output as sfh1y = sfh1g = sfh1b = sfh1c = sfh1m = 0.

The synthesizing unit 5 performs the calculation shown in the following formula based on the selected post-conversion hue area data sfh1r, sfh1g, sfh1b, sfh1c, sfh1m, sfh1y and the minimum value α, thereby generating the second color data Ro, Go, Bo is calculated.
Ro = sfh1r + sfh1m + sfh1y + α ... Formula (5-1)
Go = sfh1g + sfh1y + sfh1c + α ... Formula (5-2)
Bo = sfh1b + sfh1c + sfh1m + α… (5-3)
The internal structure of the synthesizing means 5 is the same as that of the image processing apparatus in the first embodiment shown in FIG.

  Hereinafter, the operation of the image processing apparatus according to the present embodiment will be described. Here, it is assumed that the first color data Ri, Gi, Bi shown in FIG. 13 is input to the image processing apparatus according to the present embodiment shown in FIG. At this time, the frequency characteristic conversion unit 4a performs noise removal processing on the hue region data h1r, and the frequency characteristic conversion units 4b to 4f do not perform noise removal processing, and the input hue region data h1g, h1b, h1y, h1m. , H1c are output as they are. Here, when the data selection means 11 does not select the converted hue area data based on the identification code S1, the second color data Ro is kr = 1 of the second color data Ro shown in FIG. It becomes that. In this case, in the part before the pixel position 42 that should be gray, the converted hue area data fh1r that should be zero becomes non-zero by the noise removal process, thereby causing color leakage.

  FIG. 19 is a diagram showing the second color data Ro when the data selection means 11 selects the converted hue area data based on the identification code S1. In this case, the identification code after pixel position 42 is S1 = 6 indicating a red hue, and the identification code before pixel position 42 is S1 = 12 indicating gray. For this reason, since the converted hue area data is output as sfh1y = sfh1r = sfh1g = sfh1b = sfh1c = sfh1m = 0 before the pixel position 42, color leakage that has occurred in the part before the pixel position 42 is eliminated.

  As described above, according to the image processing apparatus of the present embodiment, non-zero in the region that should be zero based on the value of the identification code S1 representing the hue information of the first color data Ri, Gi, Bi. Since the converted hue area data is removed, it is possible to avoid “color leakage” to other hues.

It is a block diagram which shows the structure of the image processing apparatus by Embodiment 1 of this invention. 3 is a block diagram illustrating an example of a configuration of hue area data calculation means 3 in the image processing apparatus according to Embodiment 1. FIG. 4 is a block diagram illustrating an example of a configuration of a frequency characteristic conversion unit 4a in the image processing apparatus according to Embodiment 1. FIG. 3 is a block diagram illustrating an example of a configuration of a combining unit 5 in the image processing apparatus according to Embodiment 1. FIG. It is a figure which shows typically the relationship between a hue and 1st chromatic color component data. It is a figure which shows typically the relationship between a hue and hue area | region data. It is a figure which shows typically the relationship between a hue and 2nd chromatic color component data. It is a figure showing an example of the 1st color data. It is a figure showing an example of the 2nd color data. It is a block diagram which shows the structure of the image processing apparatus by Embodiment 2 of this invention. It is a block diagram which shows the structure of the image processing apparatus by Embodiment 3 of this invention. FIG. 10 is a block diagram illustrating an example of a configuration of a combining unit 5b in an image processing device according to Embodiment 3. It is a figure showing an example of the 1st color data. It is a figure showing an example of the 2nd color data. It is a block diagram which shows the structure of the image processing apparatus by Embodiment 4 of this invention. It is a figure which shows the relationship between identification code | symbol S1 and 1st color data. It is a figure which shows the relationship between the value of identification code | symbol S1, and a hue. It is a figure showing an example of the relationship between identification code | cord | chord S1 and the selected 2nd hue area | region data. It is a figure showing an example of the 2nd color data.

Explanation of symbols

1, 1a αβ calculating means, 2 chromatic color component data calculating means, 3 hue area data calculating means, 4a to 4g frequency characteristic converting means, 5 and 5b combining means, 6a to 6f minimum value selecting means, and 7a to 7h data storage unit , 8 Data shift means, 9, 9a to 9f Weighted addition means, 10a to 10f Addition means, 11 Data selection means

Claims (18)

In an image processing apparatus for converting frequency characteristics of first color data representing a color image and outputting second color data corresponding to the first color data,
Hue area data calculating means for calculating first hue area data effective for a plurality of specific hue components in the color image represented by the first color data, using the first color data;
Frequency characteristic conversion means for outputting the second hue area data by independently converting the frequency characteristics of the first hue area data for each of the hue components;
An image processing apparatus comprising: combining means for calculating the second color data using the second hue area data. The hue area data calculating means calculates first hue area data effective for each of the hue components of red, green, blue, yellow, cyan, and magenta in the color image represented by the first color data. The image processing apparatus according to claim 1, wherein: Hue region data calculation means is chromatic color component data representing the size of each of the red, green, blue, yellow, cyan, and magenta color components obtained by removing the achromatic color component from the color represented by the first color data. r, g, b, y, m, c are calculated, and using the chromatic color component data, the first hue area data h1r = effective for each hue component of red, green, blue, yellow, cyan, magenta min (y, m), h1g = min (c, y), h1b = min (m, c), h1y = min (r, g), h1m = min (g, b), h1c = min (g, b The image processing apparatus according to claim 2, wherein: The frequency characteristic conversion means independently converts the frequency characteristics of the first hue area data h1r, h1g, h1b, h1y, h1c, and h1m effective for the hue components of red, green, blue, yellow, cyan, and magenta. The obtained second hue area data fh1r, fh1g, fh1b, fh1y, fh1c, fh1m are output,
The synthesizing means uses the achromatic color data α representing the achromatic color component in the color image represented by the second hue area data and the first color data, and each of red, green, and blue according to the following formula: 4. The image processing apparatus according to claim 2, wherein second color data composed of color data Ro, Go, Bo representing a component size is calculated.
Ro = fh1r + fh1m + fh1y + α
Go = fh1g + fh1y + fh1c + α
Bo = fh1b + fh1c + fh1m + α The composition means calculates the second color data using a result obtained by weighting and adding the first hue area data and the second hue area data for each hue component. The image processing apparatus according to any one of the above. Means for outputting second achromatic color data by converting frequency characteristics of the first achromatic color data representing the achromatic color component in the color image represented by the first color data;
The composition means outputs the second color data using the second hue area data and the second achromatic color data, 6. The method according to claim 1, wherein the composition means outputs the second color data. Image processing apparatus. Means for outputting an identification code indicating information relating to the hue of the color represented by the first color data;
Selecting means for selecting second hue area data relating to the hue of the color represented by the first color data based on the identification code;
The composition means outputs the second color data using the second hue area data selected by the selection means. 7. The method according to claim 1, wherein the composition means outputs the second color data. Image processing device. The image processing apparatus according to claim 1, wherein the frequency characteristic conversion unit attenuates or removes a high-frequency component of hue region data effective for a specific hue component. The image processing apparatus according to claim 1, wherein the frequency characteristic conversion unit emphasizes a contour component of hue area data effective for a specific hue component. In an image processing method for converting frequency characteristics of first color data representing a color image and outputting second color data corresponding to the first color data,
Using the first color data, calculating first hue area data effective for a plurality of specific hue components in the color image represented by the first color data,
The second hue region data is output by converting the frequency characteristics of the first hue region data independently for each of the hue components,
An image processing method, wherein the second color data is calculated using the second hue area data. 11. The first hue area data effective for each of the hue components of red, green, blue, yellow, cyan, and magenta in the color image represented by the first color data is calculated. The image processing method as described. Chromatic color component data r, g, b, y representing the size of each of the red, green, blue, yellow, cyan, and magenta color components of the color represented by the first color data excluding the achromatic color component. , M, c are calculated, and using the chromatic color component data, the first hue area data h1r = min (y, m), which is effective for each hue component of red, green, blue, yellow, cyan, and magenta. h1g = min (c, y), h1b = min (m, c), h1y = min (r, g), h1m = min (g, b), h1c = min (g, b) The image processing method according to claim 11. A second hue obtained by independently converting the frequency characteristics of the first hue area data h1r, h1g, h1b, h1y, h1c, and h1m effective for each hue component of red, green, blue, yellow, cyan, and magenta. Using the region data fh1r, fh1g, fh1b, fh1y, fh1c, fh1m, and achromatic color data α representing the achromatic color component in the color image represented by the first color data, red, green, blue 13. The image processing method according to claim 11, wherein color data of second color data including color data Ro, Go, Bo representing the size of each component is calculated.
Ro = fh1r + fh1m + fh1y + α
Go = fh1g + fh1y + fh1c + α
Bo = fh1b + fh1c + fh1m + α 13. The second color data is calculated using a result obtained by weighting and adding the first hue area data and the second hue area data for each hue component. An image processing method described in 1. By converting the frequency characteristics of the first achromatic color data representing the achromatic color component in the color image represented by the first color data, the second achromatic color data is output,
15. The image processing method according to claim 10, wherein the second color data is output using the second hue area data and the second achromatic color data. . Outputting an identification code indicating information relating to the hue of the color represented by the first color data;
Based on the identification code, select second hue area data related to the hue of the color represented by the first color data,
16. The image processing method according to claim 10, wherein the second color data is output using the selected second hue area data. 17. The image processing method according to claim 10, wherein the frequency characteristic conversion unit attenuates or removes a high-frequency component of hue region data effective for a specific hue component. 17. The image processing method according to claim 10, wherein the frequency characteristic conversion unit emphasizes a contour component of hue region data effective for a specific hue component.

Images