WO2006106750A1

WO2006106750A1 - Image processing device, image processing method, and image processing program

Info

Publication number: WO2006106750A1
Application number: PCT/JP2006/306502
Authority: WO
Inventors: Masato Tsukada; Akira Inoue; Johji Tajima
Original assignee: Nec Corporation
Priority date: 2005-03-30
Filing date: 2006-03-29
Publication date: 2006-10-12
Also published as: JPWO2006106750A1; JP4375580B2

Abstract

Image data on a color image can be converted into image data representing a vivid sketchy image with no manual operation. An original image (1) which is image data represented by a three primary color expression using RGB is inputted into first color conversion means (2), which converts the original image (1) into a brightness component and a chromaticity component in a color space in which the chromaticity does not vary or hardly varies even if the brightness varies. Filtering means (4) carries out a space filtering of the brightness acquired by the conversion by the first color conversion means (2). Second color conversion means (5) carries out deconversion of the first color conversion means (2) by using the chromaticity acquired by the conversion by the first color conversion means (2) and the space filtering results.

Description

Specification

Image processing apparatus, image processing method, and image processing program

Technical field

The present invention relates to an image processing device, an image processing method, and an image processing program.

Background art

[0002] A technique has been proposed in which color image data captured by a digital camera or the like and converted into digital data is subjected to image processing by a computer to create a pictorial image! (Refer to) o Basically, when people draw pictures easily, they often draw pictures without fine shadows, such as photographs, and often draw outlines. For this reason, image processing that creates pictorial images often creates images with fewer colors and emphasized edges. Such processing is also performed in the technique described in Patent Document 1. Specifically, each pixel in the color image data is expressed in three colors: R (red), G (green), and B (blue). The color is divided into brightness and color difference as in YCrCb. Convert to a representation in space. Then, the lightness (Y) is binarized and the lightness is set to Ymax, Ymin to reduce the number of colors. Further, an image in which a contour line is formed by edge extraction is created, and the image is created by superimposing the image on the binary image. As a result, an image such as a photographic image can be converted into a pictorial (illustration-like) image. In the technique described in Patent Document 1, after an image is created, the sharpness of the image is set by manual operation.

[0003] Further, as a color space, for example, an HSV color space that expresses a color by brightness and chromaticity is known. RGB expression power Conversion to HSV expression is described in Non-Patent Document 2, for example.

[0004] Patent Document 1: Japanese Patent Application Laid-Open No. 2004-102819 (paragraphs 0020-0026)

Non-Patent Document 1: Alvy Ray Smith, "Color CAMUT TRANSFORM PAIRS j, ComputerGraphics, Vol. 12, pp. 12—19, 1978

Disclosure of the invention Problems to be solved by the invention

[0005] However, with the technique described in Patent Document 1, it has been difficult to automatically create a colorful sketch-like painting. In other words, the lightness binary image processing, edge extraction processing, and binary image and contour image overlay processing described in Patent Document 1 can be automatically executed as processing by a computer. It is difficult to obtain a sketch-like painting with vivid colors only by its processing. The reason for this will be described below.

[0006] FIG. 5 is an explanatory diagram showing the characteristics of the color space of the YCrCb representation in which colors are represented by lightness and color difference. As shown in Fig. 5, the range of colors that can be expressed is wide at medium brightness. However, the range of colors that can be represented becomes narrower as the brightness gets brighter or darker (ie, as the Y value rises or falls from 0). And, at the brightest limit (when Y becomes the maximum possible value), only white can be expressed, and at the darkest limit (when Y becomes the minimum possible value), only black can be expressed. . Therefore, if Ymax and Ymin in the technique described in Patent Document 1 are set near the limit where only white or black can be expressed (see Fig. 5), and Y is binarized into Ymax and Ymin, it is a color difference. For Cr and Cb, the same value as before binary processing cannot be taken. Since the possible range of Cr and Cb is narrow as shown in Fig. 5, if the original Y value is binarized to Ymax and Ymin, the Cr and Cb values decrease with the binarization process. It is inevitable to let them. For this reason, an image obtained by the technique described in Document 1 has a low color saturation and is a light image.

[0007] As a pictorial image, a sketch-like image that is brightly colored when viewed by a person is often preferred. However, the image processing described in Patent Document 1 outputs a light image. Therefore, in order to create a vivid sketch-like image, a picture-like image is obtained by the automatic image processing described in Patent Document 1, and then a person looks at the image while looking at the brightness and saturation. Complicated adjustments must be made. It is preferable to have a vivid sketch-like image without any manual processing.

Therefore, the present invention provides an image processing device, an image processing method, and an image processing program for converting image data of a color image without manual intervention into image data of a vivid sketch-like image. For the purpose. Means for solving the problem

[0009] The image processing apparatus according to the present invention does not change the chromaticity of the image data components of the color image expressed in an arbitrary color space even if the lightness changes. In this case, the first color conversion means for converting to lightness and chromaticity in a color space where the chromaticity does not converge to one point, and spatial filtering for the lightness obtained by the conversion by the first color conversion means Filtering means to be applied, and second color conversion means for converting the lightness subjected to spatial filtering and the chromaticity obtained by the conversion by the first color conversion means into image data components of a color image. It is characterized by having.

[0010] According to such a configuration, the first color conversion means does not change the chromaticity of the image data component of the color image represented in an arbitrary color space even if the brightness changes. When the color space or brightness reaches the maximum value, the chromaticity is converted to the brightness and chromaticity in the color space that does not converge to one point. There is no convergence to one point. Therefore, there is no need to manually adjust the color vividness.

[0011] For example, the first color conversion unit converts the RGB component of the image data of the RGB representation color image into brightness and chromaticity in the HSV color space, and the second color conversion unit performs spatial filtering. It may be configured to convert the lightness applied and the chromaticity obtained by the conversion by the first color conversion means into the RGB component! /.

[0012] For example, when the filtering means power kernel is set to f (i, j), when at least i and j are neighboring values between the maximum and minimum possible values of i and j, f (i , j)> 0 is satisfied, and when at least one of j is at its maximum or minimum value, f (i, j) <0 is satisfied, and depending on each i and j Satisfy the condition that the sum of f (i, j) is positive. It is a configuration that performs spatial filtering by performing a convolution operation on the lightness.

[0013] According to such a configuration, it is possible to generate a contour by emphasizing a change in lightness, and to realize a conversion that does not reflect a subtle change such as a shadow of a photographic image. In addition, the image data of a color image can be converted into image data showing a sketch-like image with vivid colors. [0014] In addition, the image processing apparatus according to the present invention has a color space or brightness in which the saturation and the hue do not change even if the lightness changes in the components of the image data of the color image expressed in an arbitrary color space. Obtained by conversion by the first color conversion means that converts the lightness, saturation, and hue in a color space where saturation and hue do not converge to a single point. Depending on the saturation value, the filtering means for performing spatial filtering on the lightness and the saturation obtained by the conversion by the first color conversion means may not change the saturation value or be larger. An emphasis processing means that executes an emphasis process that changes to a value, a lightness that has undergone spatial filtering, a saturation that has undergone the emphasis process, and a hue obtained by the conversion by the first color conversion means Image image And a second color conversion means for converting the data into a component of the data.

[0015] According to such a configuration, the first color conversion unit does not change the saturation and the hue of the image data component of the color image represented in an arbitrary color space even if the lightness changes. When color space or lightness reaches the maximum value, saturation and hue are converted to lightness, saturation, and hue in a color space that does not converge to one point, so even if spatial filtering is applied to lightness, the change in lightness As a result, saturation and hue do not converge to one point. Therefore, it is not necessary to manually adjust the color vividness.

[0016] For example, the first color conversion means converts the RGB component of the image data of the color image in RGB representation into the brightness, saturation, and hue in the HSV color space, and the second color conversion means includes the space A configuration may be employed in which the lightness subjected to filtering, the saturation obtained by performing the enhancement process, and the hue obtained by the conversion by the first color conversion means are converted into RGB components.

[0017] For example, when the filtering means power kernel is f (i, j), when at least i and j are neighboring values between the maximum and minimum values of i and j, respectively, f (i , j)> 0 is satisfied, and when at least one of j is at its maximum or minimum value, f (i, j) <0 is satisfied, and depending on each i and j Satisfy the condition that the sum of f (i, j) is positive. It is a configuration that performs spatial filtering by performing a convolution operation on the lightness.

[0018] According to such a configuration, a contour can be generated by emphasizing a change in brightness, and a conversion that does not reflect a subtle change such as a shadow of a photographic image can be realized. Therefore, the image data of the color image can be converted into image data showing a sketch-like image with vivid colors.

[0019] The enhancement processing means does not change the saturation value when the saturation value is less than a predetermined threshold value, and changes the saturation value when the saturation value exceeds the threshold value. The maximum value that can be taken or a value close to the maximum value may be changed.

[0020] According to such a configuration, it is possible to convert an image of a very vivid color such as a picture drawn by a child into image data.

[0021] The enhancement processing means may be configured such that a value determined by a function having a saturation value as a variable is a saturation value.

[0022] According to such a configuration, it is possible to convert an image into image data when the degree of saturation changes gently.

[0023] Further, in the image processing method according to the present invention, the color space or lightness in which the chromaticity does not change even if the lightness changes is the maximum value of the image data component of the color image expressed in an arbitrary color space. The first color conversion step for converting to lightness and chromaticity in a color space that does not converge to the point, and spatial filtering for the lightness obtained by the conversion in the first color conversion step. A second color conversion that converts the lightness subjected to the spatial filtering, and the chromaticity obtained by the conversion in the first color conversion step into image data components of the color image. And a step.

[0024] According to such a method, in the first color conversion step, the chromaticity does not change even if the brightness of the image data components of the color image represented in an arbitrary color space changes! When the color space or brightness reaches its maximum value, the chromaticity power ^ is converted to brightness and chromaticity in a color space that does not converge to a point. Does not converge to one point. Therefore, it becomes unnecessary to manually adjust the vividness of the color.

[0025] For example, in the first color conversion step, the RGB component of the image data of the RGB representation color image is converted into brightness and chromaticity in the HSV color space, and in the second color conversion step, spatial filtering is performed. The method may be a method of converting the lightness subjected to the above and the chromaticity obtained by the conversion in the first color conversion step into RGB components. [0026] For example, in the filtering step, when the kernel is f (i, j), and at least i and j are neighboring values between the maximum and minimum values that i and j can take, respectively. When f (i, j)> 0 is satisfied and at least one of j is at its maximum or minimum value, f (i, j) <0 is satisfied, and each i , j, spatial filtering may be performed by performing a convolution operation on the lightness that satisfies the condition that the sum of f (i, j) is positive.

[0027] According to such a method, a contour can be generated by emphasizing a change in brightness, and a conversion that does not reflect a subtle change such as a shadow of a photographic image can be realized. In addition, the image data of a color image can be converted into image data showing a sketch-like image with vivid colors.

[0028] In addition, the image processing method according to the present invention has a maximum color space or lightness in which the saturation and hue do not change even if the lightness changes in the image data components of the color image represented in an arbitrary color space. The first color conversion step that converts the lightness, saturation, and hue in the color space where the saturation and hue do not converge to one point when the value is reached, and the conversion in the first color conversion step For the filtering step for performing spatial filtering on the brightness obtained in step 1 and the saturation obtained by the conversion in the first color conversion step, the saturation value is set according to the saturation value. An emphasis processing step that executes an emphasis process that does not change or changes to a larger value, a lightness that has undergone the spatial filtering, a saturation that has undergone the emphasis process, and a first color conversion step conversion And a second color conversion step of converting the hue obtained in step 1 into a component of image data of the color image.

[0029] According to such a method, in the first color conversion step, the saturation and the hue of the image data component of the color image represented in an arbitrary color space change even if the lightness changes. Yeah! When the color space or brightness reaches the maximum value, the saturation and hue are converted into brightness, saturation, and hue in a color space that does not converge to one point. The saturation and hue will not converge to one point with the change. Therefore, it is not necessary to manually adjust the color vividness.

[0030] For example, in the first color conversion step, the RGB data of the RGB color image data is generated. Are converted into lightness, saturation, and hue in the HSV color space, and in the second color conversion step, the lightness subjected to spatial filtering, the saturation subjected to enhancement processing, and the first The hue obtained by the conversion in the color conversion step may be converted to an RGB component.

[0031] For example, in the filtering step, when the kernel is f (i, j), and at least i and j are neighboring values between the maximum and minimum values that i and j can take, respectively. When f (i, j)> 0 is satisfied and at least one of j is at its maximum or minimum value, f (i, j) <0 is satisfied, and each i , j, spatial filtering may be performed by performing a convolution operation on the lightness that satisfies the condition that the sum of f (i, j) is positive.

[0032] According to such a method, a contour can be generated by emphasizing a change in brightness, and a conversion that does not reflect a subtle change such as a shadow of a photographic image can be realized. In addition, the image data of a color image can be converted into image data showing a sketch-like image with vivid colors.

[0033] In the emphasis processing step, the saturation value is not changed when the saturation value is less than a predetermined threshold value, and when the saturation value exceeds the threshold value, the saturation value is changed. It may be a method of changing to a maximum value that the saturation level can take or a value close to the maximum value.

[0034] According to such a method, it is possible to convert an image of a very vivid color such as a picture drawn by a child into image data.

[0035] In the emphasis processing step, a value determined by a function having the saturation value as a variable may be used as the saturation value.

[0036] According to such a method, it is possible to convert an image into image data when the degree of saturation changes gradually.

[0037] In addition, the image processing program according to the present invention allows a computer to store color image data components represented in an arbitrary color space in a color space that does not change even if the brightness changes. Is obtained by the first color conversion processing that converts to the lightness and chromaticity in the color space that does not converge to the point when the lightness reaches the maximum value, and the conversion by the first color conversion processing. brightness Filtering processing that applies spatial filtering to the image, and the brightness that has been subjected to spatial filtering and the chromaticity obtained by the conversion in the first color conversion processing are converted into image data components of a color image The second color conversion process is executed.

[0038] According to such a program, the color space or brightness of the image data component of the color image represented in an arbitrary color space is not changed even if the brightness changes. When the maximum value is reached, the first color conversion process is performed to convert the lightness and chromaticity in a color space where the chromaticity does not converge to one point. As a result, the chromaticity does not converge to one point. Therefore, it is not necessary to manually adjust the color vividness.

[0039] For example, by causing the computer to execute a process of converting the RGB component of the image data of the RGB-represented color image into brightness and chromaticity in the HSV color space in the first color conversion process, the second color In the conversion process, the program may execute a process for converting the lightness subjected to spatial filtering and the chromaticity obtained by the conversion in the first color conversion process into RGB components.

[0040] For example, if the kernel is set to f (i, j) by filtering processing on a computer, at least i and j are values near the maximum and minimum values that i and j can take. At some point, f (i, j)> 0 is satisfied, and when at least one of j is at its maximum or minimum value, f (i, j) <0 is satisfied. It may be a program that executes a process that performs spatial filtering by performing a convolution operation on the lightness that satisfies the condition that the sum of f (i, j) according to i and j is positive. .

[0041] According to such a program, it is possible to generate a contour by emphasizing a change in lightness, and to realize a conversion that does not reflect a subtle change such as a shadow of a photographic image. The image data of the color image can be converted into image data showing a sketch-like image with vivid colors.

[0042] Further, an image processing program according to the present invention allows a computer to use a color space or an image data component of a color image expressed in an arbitrary color space in which the saturation and hue do not change even if the lightness changes. Saturation and hue strength when the brightness reaches the maximum value ^ The first color conversion process that converts to lightness, saturation, and hue in a color space that does not converge to the point. Depending on the saturation value, the filtering process that performs spatial filtering on the lightness obtained by the conversion in the color conversion process, and the saturation obtained by the conversion in the first color conversion process, The emphasis process that does not change the saturation value, or changes it to a larger value, the lightness that has undergone spatial filtering, the saturation that has undergone the emphasis process, and the conversion in the first color conversion process. A second color conversion process for converting the obtained hue into a component of image data of a color image is executed.

[0043] According to such a program, the image data component of the color image represented in an arbitrary color space is stored in the computer in a color space or brightness in which the saturation and hue do not change even if the brightness changes. When the maximum value is reached, the first color conversion process is performed to convert the lightness, saturation, and hue in a color space where the saturation and hue do not converge to a single point, so even if spatial filtering is applied to the lightness The saturation and hue do not converge to a single point as the brightness changes. Therefore, there is no need to manually adjust the color vividness.

[0044] For example, the first color conversion process causes the computer to execute a process of converting the RGB component of the image data of the RGB representation color image into brightness, saturation, and hue in the HSV color space. In the second color conversion manual process, the lightness with spatial filtering, the saturation with the enhancement process, and the hue obtained by the conversion in the first color conversion process are converted into RGB components. The program which performs the process to perform may be sufficient.

[0045] For example, if the kernel is set to f (i, j) by filtering processing on a computer, at least i and j are values near the maximum and minimum values that i and j can take. At some point, f (i, j)> 0 is satisfied, and when at least one of j is at its maximum or minimum value, f (i, j) <0 is satisfied. It may be a program that executes a process that performs spatial filtering by performing a convolution operation on the lightness that satisfies the condition that the sum of f (i, j) according to i and j is positive. .

[0046] According to such a program, it is possible to generate a contour by emphasizing a change in brightness, and to realize a conversion that does not reflect a subtle change such as a shadow of a photographic image. The image data of the color image can be converted into image data showing a sketch-like image with vivid colors. [0047] In the emphasis process, when the saturation value is less than a predetermined threshold value, the saturation value is not changed, and when the saturation value exceeds the threshold value, the saturation value is not changed. The

Also, it may be a program that executes a process of changing to the maximum value that the saturation level can take or a value close to the maximum value.

[0048] According to such a program, it is possible to convert an image of a very vivid color such as a picture drawn by a child into image data.

[0049] A program that causes a computer to execute a process of setting a value determined by a function having a saturation value as a variable as a saturation value in an emphasis process.

[0050] According to such a program, it is possible to convert an image into image data when the degree of saturation changes gently.

The invention's effect

[0051] According to the present invention, it is possible to convert image data of a color image without manual intervention into image data of a vivid sketch-like image.

Brief Description of Drawings

FIG. 1 is an explanatory diagram showing a first embodiment of an image processing device according to the present invention.

FIG. 2 is an explanatory diagram showing an HSV color space.

FIG. 3 is an explanatory diagram showing an example of a kernel f (i, j) of a convolution operation.

FIG. 4 is an explanatory view showing a second embodiment of the image processing apparatus according to the present invention.

FIG. 5 is an explanatory diagram showing the characteristics of the color space of the YCrCb representation that expresses colors by brightness and color difference.

Explanation of symbols

[0053] 1 Original image

2 First color conversion means

4 Filtering means

5 Second color conversion means

6 Converted image

31 First brightness memory

32 chromaticity memory 33 Second brightness memory

BEST MODE FOR CARRYING OUT THE INVENTION

Hereinafter, the best mode for carrying out the present invention will be described with reference to the drawings.

[0055] Embodiment 1.

FIG. 1 is an explanatory view showing a first embodiment of an image processing apparatus according to the present invention. The image processing apparatus according to the first embodiment includes a first color conversion means 2, a chromaticity memory 32, a first brightness memory 31, a filtering means 4, a second brightness memory 33, and a second brightness memory 32. Color conversion means 5 is provided. Then, the original image 1 is input to the image processing apparatus, and the image processing apparatus outputs the converted image 6.

[0056] The original image 1 is image data of a color image (photographic image) taken by a digital camera or the like. Here, it is assumed that the original image 1 is image data represented by three primary color representations using RGB. Each pixel in the original image 1 is represented by an 8-bit value for each of R (red), G (green), and B (blue). That is, R, G, and B in each pixel take values from 0 to 255. The converted image 6 is image data obtained as a result of image processing performed on the original image 1 by the image processing apparatus, and is image data indicating a pictorial image. Here, like the original image 1, the converted image 6 is assumed to be image data of three primary color representations in which each pixel is expressed by R, G, and B.

The first color conversion unit 2 receives the original image 1. Then, the first color converting means 2 converts the component of the original image 1 (the component representing the original image 1; in this example, the RGB component) into a brightness component and a chromaticity component for each pixel. . At this time, the first color conversion means 2 does not change the chromaticity (saturation and hue) even if the lightness changes, or the lightness component and the chromaticity in the amber color space where the change of the chromaticity is small. Convert to component. As a result, the first color conversion means 2 separates the original image 1 into a lightness component and a chromaticity component. Here, a color space in which the change in chromaticity is small even if the lightness changes is defined as "When the lightness reaches the maximum value that lightness can take, the chromaticity range force does not become SO (does not converge to one point). ) "Color space". In the YCrCb representation color space, when the lightness reaches the maximum value that lightness can take, the chromaticity range converges to one point, and only white can be expressed (see Fig. 5). Therefore, the color space of YCrCb expression does not correspond to “a color space with little change in chromaticity even if the brightness changes”. Similarly, L * a * b * expression This color space does not correspond to “a color space in which the change in chromaticity is small even when the brightness changes”.

[0058] Also, here, the first color conversion means 2 is described as converting each pixel of the original image 1 expressed in RGB three primary colors into a brightness component and a chromaticity component in the HSV color space. To do. The HSV color space is a color space in which the chromaticity does not change even if the lightness changes. The chromaticity component in the HSV color space includes hue and saturation. Hue, saturation, and brightness in the HSV color space are represented by H, S, and V, respectively. The first color conversion means 2 represents the hue H, saturation S, and brightness V of each pixel obtained by the conversion by, for example, an 8-bit value (0 to 255).

[0059] Image data composed of data of lightness component V is referred to as a lightness image, and image data composed of data of chromaticity components (hue H and saturation S) is referred to as a chromaticity image. The first color conversion means 2 stores the chromaticity image obtained by the conversion process in the chromaticity memory 32 and stores the lightness image in the first lightness memory 31.

The chromaticity memory 32 stores a chromaticity image obtained by the conversion process by the first color conversion means 2. The first brightness memory 31 stores the brightness image obtained by the conversion process by the first color conversion means 2.

The filtering means 4 performs a spatial filtering process on the lightness image (that is, lightness V data) stored in the first lightness memory 31. The data obtained as a result is denoted as V ′. The filtering process by the filtering means 4 will be described later. The filtering means 4 stores the data V ′ derived by the filtering process in the second brightness memory 33.

The second brightness memory 33 stores the data V derived by the filtering means 4.

The second color conversion means 5 uses the data V ′ stored in the second lightness memory 33 as the lightness component, and the chromaticity image (that is, the hue H) stored in the data V ′ and the chromaticity memory 32. And the conversion processing of the first color conversion means 2 using the data of the saturation degree S). That is, the second color conversion means 5 converts the image data represented by the brightness component and the chromaticity component in the HSV color space into image data of the three primary colors represented by RGB. The second color conversion means 5 outputs the image data obtained by this inverse conversion as a converted image 6.

[0064] In the present embodiment, the first color conversion means 2, the filtering means 4, and the second color conversion means 2, The color conversion means 5 is realized by a CPU that operates according to a program, for example. The program may be stored in advance in a storage device (not shown) provided in the image processing apparatus. Further, the chromaticity memory 32, the first lightness memory 31, and the second lightness memory 33 are realized by, for example, a storage device (not shown) included in the image processing apparatus.

Next, the operation will be described.

When the original image 1 expressed in RGB primary color representation is input, the first color conversion means 2 converts the components (R, G, B) of the original image 1 for each pixel in the HSV color space. Convert to lightness and chromaticity components. FIG. 2 is an explanatory diagram showing the HSV color space. The vertical axis (V-axis) shown in Fig. 2 represents brightness V. In the HSV color space shown in Fig. 2, saturation (sometimes called saturation) S is represented by the distance of the V-axis force. Hue H is represented by an angle rotated around the V axis from a predetermined axis that is orthogonal to the V axis (the axis that faces in the right direction in FIG. 2). For example, if red is H = 0 °, green is H = l 20 °. As shown in Figure 2, in the HSV color space, even if the brightness V changes, the hue H and the saturation S do not change.

[0066] The first color conversion means 2 performs conversion processing from the RGB component of the original image 1 to the lightness component and chromaticity component in the HSV color space using the following conversion equations. Further, as already described, this conversion process may be performed for each pixel. However, in the following conversion formulas, R, G, B, H, S, V take values from 0 to 1. Therefore, when R, G, and B are represented by 8 bits, pre-processing is performed so that the values of R, G, and B are in the range of 0 to 1, and the following conversion formula is applied: do it. In addition, when H, S, and V are represented by 8 bits, conversion may be performed so that H, S, and V obtained by the following conversion formula are represented by 8 bits.

First, the first color conversion means 2 calculates the brightness V as V = max (R, G, B). Here, max (R, G, B) represents the maximum value among the values of R, G, B. That is, the maximum value among the values of R, G, and B is the brightness V.

In addition, the first color conversion means 2 calculates X as X = min (R, G, B). Here, min (R, G, B) represents the minimum value among the values of R, G, B. That is, let X be the minimum value of R, G, and B.

[0069] Subsequently, the first color conversion means 2 calculates the saturation S as S = (V-X) ZV. More In addition, the first color conversion means 2 calculates!:, G, b by the following formula.

[0070] r = (V-R) / (V-X)

[0071] g = (V-G) / (V-X)

[0072] b = (V-B) / (V-X)

[0073] If R = V and G = X, the first color conversion means 2 calculates h as h = 5 + b. If R = V and B = X, calculate h as h = l-g. If G = V and B = X, h is calculated as h = l + r. If G = V and R = X, calculate h as h = 3-b. If B = V and R = X, calculate h as h = 3 + g. If B = V and G = X, h is calculated as h = 5−r.

[0074] Then, the first color conversion means 2 calculates the hue H as H = hZ6, using h obtained as described above.

The first conversion means 2 stores in the chromaticity memory 32 chromaticity images, that is, image data that also serves as a data marker for chromaticity components (hue H and saturation S). In addition, a brightness image, that is, image data composed of data of the brightness component V is stored in the first brightness memory 31.

Subsequently, the filtering unit 4 performs a filtering process on the lightness component V stored in the first lightness memory 31 to calculate data V ′. Filtering process 4 generates a contour by emphasizing the change in brightness, and executes a ring-type filtering process that does not reflect subtle changes such as shading in a photographic image. The filtering means 4 realizes the filtering process as described above, for example, by calculating V ′ using the arithmetic expression of the convolution operation shown below.

[0077] [Equation 1]

V '(x, y j) V (x + i, y + j)

Here, each pixel is expressed using coordinate values (X, y) of x and y coordinates. The pixel value of each pixel (X, y) included in the brightness image (that is, the brightness of each pixel) is expressed as V (x, y). In the above equation, the pixel value of the pixel (X, y) obtained by filtering on V (x, y) is V ′ (X, y). [0079] The coefficient mag in the above arithmetic expression is a coefficient that affects all filter values. Here, the filter value is a value taken by the kernel f (i, j) of the convolution operation. Use a sufficiently large value for mag. For example, a value of 32 or more may be used as mag.

[0080] The conditions that must be satisfied by the kernel f (i, j) of the convolution operation will be described with reference to FIG. FIG. 3 is an explanatory diagram showing an example of the kernel f (i, j) of the convolution operation. The kernel has a positive value (filter value) at the center of the table illustrated in Fig. 3, a negative value (filter value) at the periphery of the table, and the sum of the filter values is positive. I just need it. If it is such a car nonore, it may not be the same carne nore as carne nore f (i, j) shown in FIG.

[0081] The value of f (i, j) in the center of the table is positive, at least that i and j are values near the maximum and minimum values that i and j can take. At some point, f (i, j)> 0. In this example, the maximum value i and j can both be 2, and the minimum value i and j can all be -2. Therefore, when i and j each have an intermediate value of 0, f (0, 0) = 13 and this value is positive. Therefore, the condition that f (i, j)> 0 is satisfied (at the center of the table) when at least i and j are values in the middle of the maximum and minimum possible values of i and j, respectively. The condition that the value of f (i, j) is positive is satisfied.

[0082] The value of f (i, j) is negative at the periphery of the table, at least when i is the maximum or minimum value that i can take, or when j is the maximum value that j can take This means that f (i, j) <0 when the value is the minimum value. In other words, f (i, j) <0 when at least one of j has the maximum or minimum value. In this example, the values when at least one of j is at its maximum value (2) or minimum value (-2) are all shown on the outermost side of the table shown in FIG. The values are both negative. Therefore, the condition is that f (i, j) <0 when at least one of j has the maximum value or the minimum value (f (i, j If the value of) is negative, the condition (! /) Is also satisfied.

[0083] The sum of the filter values being positive means that the sum of the values of f (i, j) corresponding to each i and j is positive. Since the sum of the values of f (i, j) shown in FIG. 3 is 1 and is positive, the kernel illustrated in FIG. 3 also satisfies this condition.

[0084] A convolution operation is a convolution operation using a kernel that satisfies the above conditions. The convolution operation using the same kernel as the kernel f (i, j) shown in Fig. 3 is not limited.

Note that, as shown in FIG. 3, a filter in which the portion of the kernel that is not valued is isotropic is called a ring filter.

[0086] When the filtering means 4 performs the filtering process by executing the above convolution operation and calculates V 'for each pixel, the data of V' is stored in the second brightness memory 33. Remember me.

[0087] The second color conversion means 5 uses the data V 'stored in the second brightness memory 33 as the brightness component in the HSV color space, and converts the brightness component and the chromaticity component in the HSV color space to the RGB components. The conversion is performed using the following conversion formula. However, in the following conversion formulas, R, G, B, H, S, and V are assumed to take values of 0 to 1. Therefore, when H, S, and V are represented by 8 bits, pre-processing is performed so that the values of H, S, and V are in the range of 0 to 1, and the following conversion formula is applied: do it. In addition, when R, G, B (R, G, B in the converted image 6) is represented by 8 bits, it is necessary to convert R, G, B obtained by the following conversion formula to represent 8 bits. Good.

First, the second color conversion means 5 uses the hue H,

Calculate 11 as '11. Then, the second color conversion means 5 sets the integer part of h as I. Further, the second color conversion means 5 calculates F, M, N, and K by the following formula.

[0089] F = h-I

[0090] M = V- (1 -S)

[0091] N = V- (1-(S -F))

[0092] K = V- (1-(S-(1 -F)))

Then, the second color conversion means 5 determines R, G, and B as follows according to the value of I.

When 1 = 0, the second color conversion unit 5 sets (R, G, B) = (V, Κ, M). That is, R

R, G, and B are determined as = V, G = K, and B = M.

When 1 = 1, the second color conversion unit 5 sets (R, G, B) = (N, V, M). That is, R

R, G, and B are determined as = N, G = V, and B = M.

When 1 = 2, the second color conversion unit 5 sets (R, G, B) = (M, V, K). That is, R

R, G, and B are defined as = M, G = V, and B = K. When 1 = 3, the second color conversion unit 5 sets (R, G, B) = (M, N, V). That is, R, G, and B are determined with R = M, G = N, and B = V.

When 1 = 4, the second color conversion unit 5 sets (R, G, B) = (K, M, V). That is, R

R, G, and B are defined as = K, G = M, and B = V.

[0099] When 1 = 5, the second color conversion means 5 sets (R, G, B) = (V, M, N). That is, R

R, G, and B are determined with = V, G = M, and B = N.

[0100] The second color conversion means 5 performs conversion from the data V 'corresponding to the hue H, saturation S, and lightness corresponding to each pixel into R, G, B as described above. A converted image 6 is generated.

[0101] In the above example, the first color conversion means 2 is the same as the first color conversion means 2 even if the chromaticity of the component of the original image 1 represented by the RGB primary color expression is changed. It is converted into a lightness component and a chromaticity component in a color space where does not change. Then, the filtering means 4 performs a filtering process to convert the lightness V to V ′. At this time, even if the brightness V changes, the chromaticity does not change. Therefore, the chromaticity does not converge to one point with the change of the brightness. Then, the second color conversion means 5 performs reverse conversion to RGB using the data V ′ and chromaticity. Therefore, among the components converted from the RGB component of the original image 1, only the brightness V can be converted to V, and the RGB component can be converted back. Since the filtering means 4 converts only the brightness V to V ′, it can prevent the color of the converted image 6 from fading, and it is not necessary to manually adjust the strength of the color.

[0102] Further, the filtering means 4 converts the lightness V to V 'by performing a convolution operation using a kernel that satisfies the above-described conditions. Therefore, it is possible to realize a conversion that does not reflect a subtle change such as a shadow of a photographic image while generating a contour by enhancing a change in brightness. As a result, it is possible to convert image data of a color image that does not require human intervention into image data that shows a vivid sketch-like image.

[0103] Further, in the above example, the case where the component of the original image 1 is converted into the brightness component and the chromaticity component in the HSV color space (the color space in which the chromaticity does not change even if the brightness changes) is shown. However, conversion to a color space other than the HS _V color space may be performed. The first color converting means 2 uses a color space in which the components of the original image 1 have little change in chromaticity even when the brightness changes When the maximum value is reached, it may be converted into a lightness component and a chromaticity component in a color space where the chromaticity range does not converge to one point. Then, the second color converting means 5 may perform the reverse conversion. In this case, even if the lightness V is converted to V, by the filtering means 4, the change in chromaticity does not converge to a small chromaticity force ^ point. Therefore, it is possible to prevent the color of the converted image 6 from fading, and it becomes unnecessary to manually adjust the strength of the color. Therefore, it is possible to convert color image data without human intervention into image data indicating a colorful sketch-like image.

[0104] It is to be noted that an object of the present invention is to obtain image data of a colorful sketch-like image. Therefore, the lightness component converted by the first color conversion means 2 and the color space expressing the chromaticity component are the color space where the chromaticity does not change even if the lightness changes, or the lightness is the maximum that the lightness can take. Any color space that does not converge to the chromaticity range force when it reaches the value. In order to realize vivid colors, when the brightness value is high, the range of chromaticity at the minimum value that the brightness can take is sufficient if the range of chromaticity is secured. Therefore, whether or not the chromaticity range converges to one point when the lightness reaches its minimum value does not matter.

[0105] In the above example, the original image 1 described as the original image 1 is image data represented by three primary color representations using RGB is represented by components in an arbitrary color space. The image data may be processed. For example, the original image 1 may be image data of a CMYK image represented by C (cyan), M (magenta), γ (yellow), and K (black). In this case, the first color converting means 2 converts the image data of the CMYK image into image data represented by the three primary color representations of RGB via the tristimulus value XYZ and the uniform color space CIELAB. That's fine. In this conversion, for example, a general color conversion method using a linear interpolation method based on measurement data indicating the relationship between CMYK and tristimulus values XYZ can be applied. An ICC profile or the like may be used. The first color conversion means 2 may convert the image data into the lightness component and the chromaticity component after converting into the image data represented by the RGB three primary colors in this way. The second color conversion means 5 converts the image data represented by the RGB primary color representation into an image in the original color space via the tristimulus values XYZ and the uniform color space CIELAB. That's fine. At this time, ICC You can use a profile etc.

[0106] Embodiment 2.

In the first embodiment described above, the filtering process is executed only for the lightness V of the components obtained by the conversion by the first color conversion means 2, and the chromaticity (hue H and saturation S) is executed. Is not processed. On the other hand, in the second embodiment, processing is performed for the saturation S as well. Saturation degree S is enhanced according to the value of S so that the value of S is not changed or changed to a larger value.

FIG. 4 is an explanatory view showing a second embodiment of the image processing apparatus according to the present invention. The image processing apparatus according to the second embodiment includes a first color conversion means 2, a hue memory 35, a first saturation memory 34, a first brightness memory 31, an enhancement processing means 7, and a filtering Means 4, second saturation memory 36, second brightness memory 33, and second color conversion means 5 are provided. The same components as those in the first embodiment are denoted by the same reference numerals as those in FIG. 1, and detailed description thereof is omitted.

As described in the first embodiment, the first color conversion means 2 converts the components of the original image 1 into lightness components and chromaticity components for each pixel. The chromaticity component includes the hue and the saturation, but the first color conversion means 2 separates the hue and the saturation and stores them in the memory. Specifically, the first color converting means 2 stores the hue H data of each pixel obtained by the conversion process of the original image 1 in the hue memory 35 and the saturation degree S data of each pixel is stored in the first color data. Is stored in memory 34. Further, the first color conversion means 2 stores the lightness V data of each pixel in the first lightness memory 31 as in the first embodiment. In other words, in the present embodiment, the first color conversion means 2 converts the original image 1 and separates it into lightness, saturation, and hue.

[0109] Further, the first color conversion means 2 does not change the chromaticity (saturation degree and hue) even if the lightness changes! , Or changes in chromaticity (saturation and hue) to lightness and chromaticity components in the color space. As described in the first embodiment, a color space in which the change in chromaticity (saturation and hue) is small even when the lightness changes is “when the lightness reaches the maximum value that lightness can take. , A color space where the range of chromaticity (saturation and hue) does not become zero (does not converge to one point). The hue memory 35 stores the data of the hue component H obtained by the conversion process by the first color conversion means 2. The first saturation memory 34 stores data of the saturation component S obtained by the conversion process by the first color conversion means 2. The first brightness memory 31 stores data of the brightness component V obtained by the conversion process by the first color conversion means 2.

The enhancement processing means 7 performs enhancement processing that does not change the value of S or changes it to a larger value in accordance with the value of S stored in the first saturation memory 34. Let S ′ denote the saturation S after enhancement. The enhancement processing by the enhancement processing means 7 will be described later. The enhancement processing means 7 stores the saturation S ′ after the enhancement in the second saturation memory 36.

[0112] The second saturation memory 36 stores the data S 'derived by the enhancement processing means 7.

The second color conversion means 5 uses the data V ′ stored in the second lightness memory 33 as the lightness component, V ′, and the saturation component S ′ stored in the second saturation memory 35. And the hue component H data stored in the hue memory 35 are used to perform the reverse conversion process of the conversion process of the first color conversion means 2. That is, the second color conversion means 5 converts the image data represented by the brightness component and the chromaticity component in the HSV color space into image data of the three primary colors represented by RGB.

In the present embodiment, first color conversion means 2, enhancement processing means 7, filtering means 4, and second color conversion means 5 are realized by a CPU that operates according to a program, for example. . The program may be stored in advance in a storage device (not shown) included in the image processing apparatus. Also, the hue memory 35, the first saturation memory 34, the second saturation memory 36, the first brightness memory 31, and the second brightness memory 33 are, for example, storage devices (see FIG. (Not shown)).

Next, the operation will be described.

When the original image 1 expressed in RGB primary color representation is input, the first color conversion means 2 converts the original image 1 for each pixel with a lightness component V, a saturation component H, in the HSV color space. And convert to hue component H. The conversion process by the first color conversion means 2 is the same as the conversion process by the first color conversion means 2 shown in the first embodiment. The first conversion means 2 stores the hue component H data of each pixel in the hue memory 35. Similarly, saturation of each pixel The data of the degree component S is stored in the first saturation memory 34. Further, the data of the brightness component V of each pixel is stored in the first brightness memory 31.

Subsequently, the filtering means 4 performs the same filtering process as in the first embodiment, converts the data of the lightness component V into V, and stores the data V ′ in the second lightness memory 33. Let me.

[0117] Further, the enhancement processing means 7 performs enhancement processing on the saturation component S of each pixel stored in the first saturation memory 34, and uses the saturation component S 'after enhancement processing as the second saturation level. Store in memory 36.

[0118] The enhancement processing means 7 executes the enhancement processing as follows, for example. The enhancement processing means 7 reads the value of the saturation component S from the first saturation memory 34. Then, the value of the saturation component S is compared with a predetermined threshold value (denoted as th). The enhancement processing means 7 determines S ′ as S ′ = S when S is less than the threshold value th. Further, when S is equal to or greater than the threshold th, the enhancement processing means 7 determines the value of S ′ as a maximum value that S can take or a value close to the maximum value (denoted as Smax). For example, it is assumed that th = 32, and S can take a value in the range of 0 to 255. Assume that Smax = 255. Emphasis processing means 7 finds 3 ′ with S ′ = 3 if S is 32. Further, the enhancement processing means 7 sets S ′ = 255 if S ≧ 32.

[0119] In this way, when the value of S becomes equal to or greater than the threshold th, by performing enhancement processing with S ′ = Smax, the finally obtained converted image 6 is like a picture drawn by a child. It can be generated as very vivid color image data.

[0120] In the above example, the case where S '= Smax is set when the value of S is equal to or greater than the threshold value th. The value of S ′ may be determined by the function S ′ = g (S) that changes to An example of such a function g (S) is shown below. In the example shown below, it is assumed that S can take a value in the range of 0 to 255.

[0121] S, = S (when S is 32)

S '= 4- S- 96 (when 32≤S≤87)

S '= 255 (when S≥88)

[0122] When deriving S 'determined by such a function, for example, 256 types of S from 0 to 255 The value of S, corresponding to the value of, is calculated in advance. And each S value and S according to that S value

, And are stored in advance in a lookup table (not shown). The enhancement processing means 7 may derive the value of S ′ corresponding to the value of S by referring to this lookup table.

After the filtering process by the filtering means 4 and the enhancement process by the enhancement processing means 7, the second color conversion means 5 uses the data V ′ stored in the second brightness memory 33 as the brightness component, and V ′ And using the saturation component S ′ data stored in the second saturation memory 35 and the hue component H data stored in the hue memory 35, the conversion processing of the first color conversion means 2 is performed. Inverse conversion processing is performed. The conversion process by the first color conversion unit 2 is performed by the second color conversion unit 5 shown in the first embodiment, except that S ′ is used instead of the saturation component S. It is the same.

In the present embodiment, as in the first embodiment, it is possible to convert image data of a color image without human intervention into image data of a colorful sketch-like image. Also, the image data of the image obtained by converting the image data to a very vivid color image such as a picture drawn by a child, or by slightly changing the degree of saturation slightly (slowly). Can be generated.

[0125] Also, as in the first embodiment, the original image 1 may be image data represented by components in an arbitrary color space. For example, it may be image data of a CMYK image. In this case, the operations of the first color conversion means 2 and the second color conversion means 5 are the same as those described in the first embodiment.

Industrial applicability

The present invention can be applied to uses such as a projector color correction apparatus, for example. For example, the present invention can be applied to a program for realizing a projector color correction function by a computer.

Claims

The scope of the claims

[1] Color image image components expressed in an arbitrary color space are converged to a color space where the chromaticity does not change even if the brightness changes or when the brightness reaches the maximum value ^ A first color conversion means for converting to lightness and chromaticity in a color space;

Filtering means for performing spatial filtering on the brightness obtained by the conversion by the first color conversion means;

Second color conversion means for converting the lightness subjected to the spatial filtering and the chromaticity obtained by the conversion by the first color conversion means into image data components of the color image; An image processing apparatus.

[2] The first color conversion means converts the RGB component of the image data of an RGB representation color image into brightness and chromaticity in the HSV color space,

2. The image processing device according to claim 1, wherein the second color conversion means converts the lightness subjected to spatial filtering and the chromaticity obtained by the conversion by the first color conversion means into RGB components. .

[3] When the kernel is f (i, j), the filtering means uses f (i) when at least i and j are neighboring values between the maximum and minimum possible values of i and j, respectively. i, j)> 0, and at least one of j is at its maximum or minimum value, f (i, j) satisfies 0 and depends on each i and j 3. The image processing device according to claim 1, wherein spatial filtering is performed by performing a convolution operation that satisfies a condition that a sum of f (i, j) is positive on the lightness.

[4] The image data components of a color image expressed in an arbitrary color space are combined with the saturation level and the saturation level when the saturation or hue does not change even if the brightness changes or the brightness reaches the maximum value. A first color conversion means for converting lightness, saturation, and hue in a color space in which the hue does not converge to one point;

For the saturation obtained by the conversion by the first color conversion means, emphasis processing is performed in which the saturation value is not changed or is changed to a larger value depending on the saturation value. Emphasis processing means to perform,

The lightness that has been subjected to the spatial filtering, the saturation that has been subjected to the enhancement processing, and the hue obtained by the conversion by the first color conversion means are converted into image data components of the color image. And a second color conversion means

An image processing apparatus.

[5] The first color conversion means converts the RGB component of the image data of the RGB representation color image into brightness, saturation, and hue in the HSV color space,

The second color conversion means converts the lightness subjected to spatial filtering, the saturation degree subjected to the enhancement processing, and the hue obtained by the conversion by the first color conversion means into RGB components. The image processing apparatus according to claim 4.

[6] When the kernel is set to f (i, j), the filtering means uses f (i) when at least i and j are neighboring values between the maximum and minimum values i and j can take. i, j)> 0, and at least one of j is at its maximum or minimum value, f (i, j) satisfies 0 and depends on each i and j Spatial filtering is performed by performing a convolution operation on the lightness that satisfies the condition that the sum of f (i, j) is positive.

The image processing device according to claim 4 or 5.

[7] The emphasis processing means does not change the saturation value when the saturation value is less than a predetermined threshold value, and saturates the saturation value when the saturation value exceeds the threshold value. Change to the maximum value that can be taken by the degree or a value close to the maximum value

The image processing apparatus according to any one of claims 6 and 6.

8. The image processing device according to claim 4, wherein the enhancement processing means uses a value determined by a function having a saturation value as a variable as a saturation value.

[9] Color image image components expressed in an arbitrary color space are converged to a color space where the chromaticity does not change even if the lightness changes or when the lightness reaches the maximum value ^ point A first color conversion step for converting to lightness and chromaticity in a color space;

Spatial filtering is applied to the lightness obtained by the conversion in the first color conversion step. The lightness subjected to the spatial filtering and the lightness obtained by the conversion in the first color conversion step. And a second color conversion step for converting the obtained chromaticity into image data components of the color image.

An image processing method.

[10] In the first color conversion step, the RGB component of the image data of the RGB representation color image is converted to lightness and chromaticity in the H SV color space.

The lightness that has undergone spatial filtering in the second color conversion step and the chromaticity obtained by the conversion in the first color conversion step are converted into RGB components.

The image processing method according to claim 9.

[11] In the filtering step, if f (i, j) is the kernel, at least i and j are neighboring values between the maximum and minimum possible values of i and j, respectively. i, j)> 0, and when i or j is at its maximum or minimum value, f (i, j) <0 is satisfied, depending on each i, j Perform spatial filtering by performing a convolution operation on the lightness that satisfies the condition that the sum of f (i, j) is positive.

The image processing method according to claim 9 or 10.

[12] The image data components of a color image expressed in an arbitrary color space are the saturation and saturation values when the saturation or hue does not change even when the brightness changes or the brightness reaches the maximum value. And a first color conversion step for converting to lightness, saturation, and hue in a color space where the hue does not converge to one point;

Spatial filtering is applied to the lightness obtained by the conversion in the first color conversion step, and the saturation obtained by the conversion in the first color conversion step is determined according to the value of the saturation. An emphasis processing step for executing an emphasis process that does not change the saturation value or changes the saturation value to a larger value;

The brightness that has been subjected to the spatial filtering, the saturation that has been subjected to the enhancement processing, and the hue obtained by the conversion in the first color conversion step are used as image data components of the color image. A second color conversion step to convert

An image processing method.

[13] In the first color conversion step, the RGB component of the image data of the RGB color image is convert to lightness, saturation and hue in sv color space,

In the second color conversion step, the lightness that has undergone spatial filtering, the saturation that has undergone enhancement processing, and the hue obtained by the conversion in the first color conversion step are converted into RGB components. Do

The image processing method according to claim 12.

[14] In the filtering step, if f (i, j) is the kernel, at least i and j are neighboring values between the maximum and minimum possible values of i and j, respectively. i, j)> 0, and when i or j is at its maximum or minimum value, f (i, j) <0 is satisfied, depending on each i, j Perform spatial filtering by performing a convolution operation on the lightness that satisfies the condition that the sum of f (i, j) is positive.

The image processing method according to claim 12 or claim 13.

[15] In the emphasis processing step, when the saturation value is less than a predetermined threshold value, the saturation value is not changed, and when the saturation value exceeds the threshold value, the saturation value is changed. Change to the maximum possible value of saturation or a value close to the maximum

The image processing method according to any one of claims 12 to 14.

[16] In the emphasis processing step, the value determined by the function with the saturation value as a variable is the saturation value.

The image processing method according to any one of claims 12 to 14.

[17] On the computer,

A color space in which the image data component of a color image expressed in an arbitrary color space does not change even if the brightness changes, or the chromaticity power ^ does not converge to a point when the brightness reaches the maximum value The first color conversion process to convert to lightness and chromaticity in

Filtering processing for performing spatial filtering on the brightness obtained by the conversion in the first color conversion processing; and

Causing a second color conversion process to convert the lightness subjected to the spatial filtering and the chromaticity obtained by the conversion in the first color conversion process into image data components of the color image; Image processing program.

[18] On the computer, In the first color conversion process, the RGB component of the image data of the RGB representation color image is converted to lightness and chromaticity in the HSV color space.

In the second color conversion process, execute a process of converting the lightness subjected to spatial filtering and the chromaticity obtained in the conversion in the first color conversion process into RGB components.

The image processing program according to claim 17.

[19] On the computer,

In the filtering process, when f (i, j) is the kernel, at least i and j are f (i, j) when they are near the maximum and minimum possible values of i and j, respectively. )> 0, and when at least one of j is at its maximum or minimum value, f (i, j) satisfies 0 and f ( 19. The image processing program according to claim 17, wherein a process of applying spatial filtering is executed by performing a convolution operation that satisfies the condition that the sum of i, j) is positive on the lightness.

[20] On the computer,

The image data components of a color image expressed in an arbitrary color space are the same as the saturation and hue when the saturation or hue does not change even if the brightness changes, or when the brightness reaches the maximum value. A first color conversion process that converts to lightness, saturation, and hue in a color space that does not converge to one point,

A filtering process for applying spatial filtering to the brightness obtained by the conversion in the first color conversion process;

The enhancement processing for changing the saturation value to a larger value or not depending on the saturation value with respect to the saturation obtained by the conversion in the first color conversion processing, and the spatial filtering And a hue obtained by the conversion in the first color conversion process into a component of image data of the color image. Color conversion process

An image processing program for executing

[21] On the computer,

In the first color conversion process, the RGB component of the image data of the RGB representation color image is converted to lightness, saturation, and hue in the HSV color space. In the second color conversion manual processing, the brightness that has been spatially filtered, the saturation that has been subjected to the enhancement processing, and the hue obtained by the conversion in the first color conversion processing are converted into RGB components. Execute the process to convert

The image processing program according to claim 20.

[22] On the computer,

In the filtering process, when f (i, j) is the kernel, at least i and j are f (i, j) when they are near the maximum and minimum possible values of i and j, respectively. )> 0, and when at least one of j is at its maximum or minimum value, f (i, j) satisfies 0 and f ( The image processing program according to claim 20 or 21, wherein a process of performing spatial filtering is executed by performing a convolution operation that satisfies the condition that the sum of i, j) is positive on the lightness.

[23] On the computer,

In the emphasis process, when the saturation value is less than a predetermined threshold value, the saturation value is not changed, and when the saturation value exceeds the threshold value, the saturation value can be taken. Execute processing to change to the maximum value or a value close to the maximum value

The image processing program according to any one of claims 20 to 22.

[24] On the computer,

In the emphasis process, execute the process of setting the value determined by the function with the saturation value as a variable to the saturation value.

The image processing program according to any one of claims 20 to 22.