JP2005174133A - Image evaluation device, image evaluation method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform evaluation considering a human visual characteristic, in quality evaluation of a gradation image. <P>SOLUTION: An image input means 101 reads an evaluated image. and inputs image data showing a position of a pixel and a pixel value of the evaluated image. A color information conversion means 102 converts the pixel value into color information showing one of brightness, chroma, hue, density, luminance, chromaticity, and intensity. A spatial frequency information conversion means 103 corrects the color information by the human visual spatial frequency characteristic. An image characteristic amount calculation means 104 finds a difference between adjacent pixels of the color information corrected by the spatial frequency information conversion means 103. An evaluation value calculation means 105 compares the difference found by the image characteristic amount calculation means 104 with a threshold value of a tone jump amount which a person can perceive to extract a tone jump having the tone jump amount which the person can perceive, and finds the tone jump amount which the person actually perceive for the extracted tone jump. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a technique for evaluating the quality of a color image.

  Generally, in an image forming apparatus such as a printer or a copying machine, tone reproducibility is often evaluated through output and analysis of a tone image. Here, the gradation image refers to an image in which at least one of color information such as brightness, saturation, and hue changes continuously. Originally, when such a gradation image is output to an ideal printer, the output result should reproduce the gradation smoothly. However, an actual printer may output an image with a distorted gradation such as a gradation skip or streaky unevenness. This disturbance occurs in the same way when a normal image is output, and causes a significant decrease in printer performance. Therefore, quantitatively evaluating the disturbance of the gradation image is very important in evaluating the drawing performance of the image forming apparatus. The same applies to an image display device such as a CRT (Cathode Ray Tube).

  As a method for analyzing and evaluating a gradation image, for example, human visual comparison evaluation is widely performed. This method is called so-called subjective evaluation, and evaluates the quality of a gradation image and the degree of damage by subjective judgment through human vision. However, such subjective evaluation has a problem that evaluation results vary by the evaluator and it takes time and effort.

In order to solve such problems, in recent years, a technique has been proposed for objective evaluation that does not depend on human subjective judgment on a gradation image output by an image forming apparatus or the like. (For example, Patent Documents 1 to 5).
In the technique described in Patent Document 1, a difference in optical information between adjacent gradation levels is calculated using at least one of the brightness or density of the output patch image, and the calculated difference and human perception The gradation image is evaluated by comparing with a threshold value of gradation skip that can be performed. Also, in order to reduce errors due to in-plane variations, patch images are used in which adjacent gradation level patches are arranged spatially adjacent to each other.

In the technique described in Patent Document 2, a difference in optical information between adjacent gradation levels is calculated using at least one optical information among the density, brightness, and chromaticity of a patch image.
In the technique described in Patent Document 3, a gradation image is read for each small region, and if the color difference between adjacent small regions is equal to or less than a threshold that can be perceived by humans, both are regarded as regions of the same gradation, and the color difference has a threshold. If it exceeds, it is regarded as a region of a different gradation. Then, an evaluation value is calculated using a dispersion value and an average value of color differences between adjacent gradation regions.

In Patent Documents 4 and 5, a curve representing a change in optical information (for example, brightness) obtained by measuring a gradation image is smoothed (may be obtained by approximation or interpolation), and the smoothed curve and the original are obtained. The gradation image is evaluated by obtaining the difference from the curve.
Japanese Patent Laid-Open No. 11-25274 JP-A-11-39486 JP 2001-128017 A JP 2002-86705 A JP 2003-16443 A

However, with the technique described in the above-mentioned patent document, it is difficult to correctly evaluate the gradation image for the following reason.
In the case of a color image, the brightness (density) may be the same even though the colors are different. In such a case, in order to evaluate the difference in color, it is necessary to evaluate using saturation or hue. In the technique described in Patent Document 1, since evaluation is performed using either brightness or density, it is difficult to appropriately evaluate a color image. Further, although patch images are used as the images to be evaluated, each patch outputs a uniform color and does not have a continuous gradation change. Accordingly, the image quality degradation peculiar to the gradation image such as gradation skip and gradation unevenness does not appear, and therefore it is not suitable for evaluation of the gradation image.
Since Patent Document 2 also uses a patch image, it has the same problem as described above.

  In the technique described in Patent Document 3, a comparison is made with a variance value having a width exceeding a threshold value, but this variance value is different from an actual evaluation value. For example, in the case of a printer, since color correction is performed on input, each printer has a unique gradation curve. For example, since the human eye is sensitive to changes from highlight to halftone, this tone curve is adjusted to reduce the color gamut compression between the highlight portion and the intermediate portion. For this reason, even if the gradation image looks natural to humans, the variance value may increase. Further, when the gradation is expressed by the area ratio of the halftone dots, the variation of the measured gradation becomes large due to the influence of the screen, and the threshold value may be exceeded. Also, if the reading range is increased in order to reduce the influence of the screen, it is not possible to evaluate gradation skip.

  The techniques described in Patent Documents 4 and 5 have a problem that the evaluation result depends on the smoothed curve. Further, although the evaluation is performed by the first-order differential of the difference, the absolute value of the first-order differential may be increased even if the difference is small. Thereby, a large noise may occur in the evaluation value. Further, when the difference is a roof edge shape, it is inconspicuous in the visual evaluation, but the portion where the absolute value of the first derivative increases is increased, so that it is difficult to obtain an evaluation result consistent with the subjective evaluation. Further, it is difficult to evaluate a two-dimensional gradation image by the above method.

  The present invention has been made under the above-described background, and an object of the present invention is to provide a technique that can perform evaluation in consideration of human visual characteristics in quality evaluation of a gradation image.

  In order to solve the above-described problems, the present invention provides an image input unit that reads an image to be evaluated and inputs image data representing a pixel value and a pixel position of the image to be evaluated; Color information conversion means for converting color information representing any one of degree, hue, density, luminance, chromaticity, and intensity; spatial frequency information conversion means for correcting the color information according to human visual spatial frequency characteristics; and An image feature quantity calculating means for obtaining a difference between adjacent pixels of the color information corrected by the spatial frequency information converting means, a difference obtained by the image feature quantity calculating means, and a threshold value for a gradation skip amount perceivable by a human. It has an evaluation value calculation means for extracting a gradation skip having a gradation skip amount that can be perceived by humans by comparing, and obtaining a gradation jump amount that is actually perceived by the human with respect to the extracted gradation skip Providing an image evaluation device.

  Further, the present invention provides an image input means for reading an image to be evaluated and inputting image data representing a pixel value and a pixel position of the image to be evaluated, and the pixel value for brightness, saturation, hue, density, and luminance. Color information conversion means for converting into plural kinds of color information of chromaticity and intensity, spatial frequency information conversion means for correcting each of the color information by spatial frequency characteristics of human vision, and the spatial frequency information conversion By comparing an image feature amount calculating means for obtaining a difference between adjacent pixels of color information corrected by the means, and a difference obtained by the image feature amount calculating means and a threshold value of a gradation jump amount perceivable by a human An evaluation value calculating means for extracting a gradation skip having a gradation skip amount that can be perceived by a human and obtaining a gradation skip amount actually perceived by a human with respect to the extracted gradation skip; and the evaluation value calculating means To provide an image evaluation device having a total evaluation value calculating means for summing by weighting based on each of the obtained gradation jump amount to the ratio of contribution to human perception.

  The present invention also includes an image input step of reading an image to be evaluated and inputting image data representing a pixel value and a pixel position of the image to be evaluated, and the pixel value as a value, brightness, saturation, hue, density, and luminance. A color information conversion step for converting the color information into color information representing any one of chromaticity and intensity, a spatial frequency information conversion step for correcting the color information by a spatial frequency characteristic of human vision, and correction by the spatial frequency information conversion means An image feature amount calculating step for obtaining a difference between adjacent pixels in the color information obtained by comparing the difference obtained by the image feature amount calculating unit with a threshold value of a gradation jump amount that can be perceived by a human. An image having an evaluation value calculation step for extracting a gradation skip having a possible gradation skip amount and obtaining a gradation skip amount actually perceived by a human with respect to the extracted gradation skip To provide an evaluation method.

  Further, the present invention provides an image input means for reading an image to be evaluated and inputting image data representing a pixel value and a pixel position of the image to be evaluated, and the pixel value for brightness, saturation, hue. Color information conversion means for converting color information representing any one of density, luminance, chromaticity, and intensity, spatial frequency information conversion means for correcting the color information according to human visual spatial frequency characteristics, and the spatial frequency information Comparing an image feature amount calculation means for obtaining a difference between adjacent pixels of color information corrected by the conversion means, and a difference obtained by the image feature amount calculation means and a threshold value of a gradation jump amount perceivable by humans This function functions as an evaluation value calculation means that extracts gradation skips with a gradation skip amount that can be perceived by humans, and calculates the amount of gradation skips that humans actually perceive for the extracted gradation skips. To provide because of the program.

  According to the present invention, the image input means reads the image to be evaluated and inputs image data representing the pixel value and the pixel position of the image to be evaluated. Color information conversion means converts the pixel value into color information representing any one of brightness, saturation, hue, density, luminance, chromaticity, and intensity. Spatial frequency information conversion means corrects the color information by the spatial frequency characteristics of human vision. The image feature amount calculating means obtains a difference between adjacent pixels of the color information corrected by the spatial frequency information converting means. The evaluation value calculating means extracts a gradation skip having a gradation skip amount that can be perceived by humans by comparing the difference obtained by the image feature amount calculating unit with a threshold value of the gradation skip amount that can be perceived by humans. Then, the amount of gradation skip that is actually perceived by the human with respect to the extracted gradation skip is obtained.

  According to the present invention, it is possible to perform an evaluation in consideration of human visual characteristics in the quality evaluation of a gradation image.

Embodiments of the present invention will be described below with reference to the drawings.
[First Embodiment]
[Constitution]
The image evaluation apparatus 10 analyzes a gradation image output by an image forming apparatus such as a printer or a copier, or a gradation image output by an image display apparatus such as a CRT, and uses the analysis result to analyze the gradation image. The evaluation value of the toned image is obtained, and thereby the drawing performance of the image forming apparatus or the image display apparatus is evaluated.

FIG. 1 is a diagram illustrating a hardware configuration of the image evaluation apparatus 10. The image evaluation apparatus 10 includes a CPU (Central Processing Unit) 11, a RAM (Random Access Memory) 12, a nonvolatile memory 13, and a communication IF (interface) 14.
A program 131 is written in the nonvolatile memory 13. When the image evaluation apparatus 10 is powered on (not shown), the CPU 11 reads the program 131 written in the nonvolatile memory 13 and executes the RAM 12 as a work area. When the CPU 11 executes the program 131, the image evaluation apparatus 10 is virtually formed with a module group shown in FIG. Note that these module groups may be mounted on hardware, and the CPU 11 may control the hardware to cause each module to function. Alternatively, the program 131 may be stored in an external device, and the program 131 may be downloaded via a communication network and stored in the nonvolatile memory 13.

An image input device 20 for inputting a gradation image (evaluated image) to be evaluated is connected to the communication IF 14 of the image evaluation device 10. Here, the image input device 20 is, for example, a one-dimensional scanning scanner (flatbed scanner). The image input device 20 is not limited to this, and other devices having a spectral sensitivity equivalent to the color matching function in the XYZ color system, that is, a microdensitometer, a color scanner, a monochrome scanner, a drum scanner, A densitometer, a colorimeter, a film scanner, or the like may be used. Moreover, you may use the general purpose image input device 20 which outputs RGB signals, such as a digital camera. When the image to be evaluated is an image displayed on the screen of an image display device such as a CRT, image data may be acquired using a luminance meter.
The to-be-evaluated image is a gradation image whose gradation changes continuously, and is a gradation image output by an image forming apparatus such as a printer or a copier, or a gradation output by an image display apparatus such as a CRT. It is an image. Further, the image to be evaluated is any one of a reflective object, a transmissive object, and a light emitting object.

Next, a module group virtually formed in the image evaluation apparatus 10 by the CPU 11 executing a program will be described.
The image input unit 101 is a unit that optically reads a color gradation image (evaluated image) to be evaluated, and inputs image data representing a pixel value and a pixel position of the evaluated image. Specifically, the image input unit 101 causes the image input device 20 to read the evaluation target image and generate image data representing the evaluation target image. Then, the generated image data is received from the image input device 20, and the received image data is stored in the nonvolatile memory 13. This image data is raster data, and position information representing the position of each pixel on the plane is held.

  The color information conversion unit 102 is a unit that converts pixel values into color information. The color information is lightness, saturation, and hue constituting the three attributes of color. The color information conversion unit 102 converts pixel values included in the image data generated by the image input unit 101 into an L * a * b * color space or L * u * v defined by CIE (Commission Internationale de I'Eclariage). * Convert to each component in the color space to obtain color information. This conversion is performed using an ICC (International Color Consortium) profile, a multiple regression equation, or a lookup table. Note that the color information may be any one of density, luminance, chromaticity, and intensity. Here, the intensity is the intensity of the image signal input by the image input device. Hereinafter, an example using lightness among the color information will be described.

The spatial frequency information conversion unit 103 is a unit that corrects the color information obtained by the color information conversion unit 102 with the spatial frequency characteristics of human vision. Specifically, the spatial frequency information conversion means 103 first converts the lightness distribution in the real space into the spatial frequency distribution in the frequency space. For this conversion, any method such as DFT (Discrete Fourier Transform), FFT (Fast Fourier Transform), and DCT (Discrete Cosine Transform) can be used as long as it can convert the distribution in the real space to the spatial frequency distribution. Good. Furthermore, the spatial frequency information conversion means 103 multiplies the spatial frequency distribution by a function VTF (Visual Transfer Function) representing the spatial frequency characteristics of human vision, and inversely transforms the spatial frequency distribution obtained by the multiplication to obtain a brightness distribution. obtain. In this way, the brightness corrected with the spatial frequency characteristics of human vision is obtained.
Note that the spatial frequency information conversion means 103 may use a method of performing convolution filtering on the real space instead of the calculation via the frequency space.

The image feature amount calculation unit 104 is a unit that obtains a difference between adjacent pixels of the color information corrected by the spatial frequency information conversion unit 103. Specifically, the second derivative of the brightness distribution corrected with the spatial frequency characteristics of human vision is calculated. For the calculation of the second derivative, the lightness distribution is filtered using a Laplacian filter. 3A and 3B show examples of Laplacian filter operators.
If there is noise generated in the image input unit 101, this noise may be overemphasized. In order to prevent this noise from being emphasized too much, as shown in FIGS. 4 (a) and 4 (b), an operator is applied to the brightness distribution averaged in the area of n × n (n is an integer). It is preferable to multiply. As an example, when observing a reflective document at a clear viewing distance of about 40 cm, the sensitivity of visual characteristics increases at about 1 cycle / mm, so an operator that performs a local product-sum operation on a 3 × 3 mm region of a two-dimensional lightness distribution. Use.

  The evaluation value calculation unit 105 extracts a gradation jump having a gradation jump amount that can be perceived by humans by comparing the difference obtained by the image feature amount calculation unit 104 with a threshold value of the gradation jump amount that can be perceived by humans. It is a means for obtaining the amount of gradation skip that is actually perceived by the human with respect to the extracted gradation skip. First, the evaluation value calculation unit 105 uses the threshold value calculation unit 106 to obtain a threshold value of a gradation skip amount that can be perceived by humans using color information. Then, by comparing the difference obtained by the image feature amount calculation unit 104 with this threshold value, a gradation skip having a gradation skip amount that can be perceived by a human is extracted. For the extracted gradation skip, the subtraction means 107, the first correction means 108, and the second correction means 109 are further used to determine the gradation skip amount actually perceived by humans.

The calculation of the threshold value by the threshold value calculation unit 106 will be described. As an example, a formula for calculating a threshold value Ld of gradation skip that can be perceived by humans in a gradation image that continuously changes from white to black is shown in Equation 1. According to this, the threshold value depends on the lightness (L *). Also, as a matter of course, the threshold expression is different for different colors.

The subtracting means 107 is a means for subtracting the amount corresponding to the threshold value from the gradation skip amount extracted using the threshold value. When the gradation skip amount is less than or equal to the threshold value, it is not regarded as a gradation skip.
The first correction unit 108 is a unit that corrects the gradation skip amount subtracted by the subtraction unit 107 by using the color information. The correction by the first correction unit 108 will be described below.
The amount of gradation skip that is actually perceived by humans when gradation jump is perceived by humans varies depending on the lightness at which the gradation skip occurred. That is, in order to obtain the gradation skip amount G ₁ felt by humans, it is necessary to correct the gradation skip amount G ₀ using the lightness L * at which the gradation skip has occurred. An example of a formula for calculating the gradation skip amount G ₁ corrected by the brightness is shown in Formula 2.

The second correction unit 109 is a unit that corrects the gradation skip amount corrected by the first correction unit 108 so as to have the same rate as human perception. The correction by the second correction unit 109 will be described below.
Not the human vision and the equal pace even tone jump amount G ₁ corrected by the first correction means 108. That is, correction by the gradation skip amount G ₁ is necessary. An example of a formula for calculating the gradation skip amount G ₂ corrected by the gradation skip amount G ₁ is shown in Equation 3.

Then, the evaluation value calculation means 105 integrates the gradation skip amounts corrected as described above, and uses the integrated value as the gradation evaluation value SL *. A formula for calculating the evaluation value SL * is shown in Equation 4.

[Operation]
The operation of the image evaluation apparatus 10 having the above configuration will be described. However, since the image evaluation apparatus 10 is an apparatus that operates by using software as hardware, in the following description, the subject of the operation is hardware, not a virtually formed module. FIG. 5 is a diagram illustrating a flow of processing performed by the CPU 11 executing the program 131. Here, it is assumed that the image evaluation apparatus 10 and the image input apparatus 20 are powered on and a program is executed in the image evaluation apparatus 10.

First, an operator places a paper on which an image to be evaluated is printed on the image input device 20 and inputs an instruction to start reading the image to be evaluated to the image evaluation device 10. Then, the image input device 20 reads the image to be evaluated, and generates image data including the obtained RGB signal and pixel position information. This image data is transmitted to the image evaluation apparatus 10 (step S01).
The image evaluation device 10 receives the image data transmitted from the image input device 20 and stores it in the nonvolatile memory 13.

Next, the CPU 11 reads out the image data stored in the nonvolatile memory 13 and converts the RGB values included in the image data into colorimetric values L *, a *, and b * in the CIE 1976 L * a * b * color space. Conversion is performed (step S02). Hereinafter, lightness (L *) will be described as an example.
Next, the CPU 11 corrects the brightness distribution using the spatial frequency characteristics of human vision (step S03). Specifically, first, the brightness distribution is converted into a spatial frequency distribution by Fourier transform. The spatial frequency distribution obtained by converting the lightness distribution is multiplied by a function representing the spatial frequency characteristics of human vision, and converted to a lightness distribution in real space by inverse Fourier transform. As a result, a brightness distribution corrected with the spatial frequency characteristics of human vision is obtained. The brightness distribution thus obtained is stored in the nonvolatile memory 13 in association with the pixel position information.

Next, the CPU 11 performs a filter process using a Laplacian filter on the lightness distribution corrected as described above, and obtains a second derivative of the lightness distribution (step S04). Thereby, the gradation skip amount of the image to be evaluated is obtained.
Next, the CPU 11 obtains an evaluation value of the gradation skip amount of the image to be evaluated (step S05). First, the CPU 11 reads the brightness distribution corrected by the human spatial frequency characteristics from the nonvolatile memory 13 and obtains a threshold value corresponding to the brightness of each pixel by using Equation (1). Then, the CPU 11 subtracts the threshold value equivalent from the second derivative of the brightness distribution for each pixel. The value after subtracting the amount corresponding to the threshold is the gradation skip that can be perceived by humans. However, if the second derivative is less than or equal to the threshold value, the subtraction value is set to zero.

The gradation skip obtained as described above is further corrected using Equations 2 and 3. Thus, the gradation skip amount actually perceived by a human is obtained.
Then, the gradation skip amounts corrected using Equation 2 and Equation 3 are integrated according to Equation 4, and the integrated value is used as an evaluation value.

As described above, in the evaluation of the gradation image, it is possible to perform evaluation in consideration of human visual characteristics.
According to the present invention, since the evaluation is performed using at least one of lightness, saturation, hue, density, luminance, chromaticity, and intensity, even if the color image has a small lightness difference, The evaluation of gradation can be performed by using the hue or the like.
Further, since the threshold value of the gradation skip amount that can be perceived by humans is obtained using the color information, it is possible to set the threshold value depending on the value of the color information.

Further, since the threshold value is subtracted from the gradation skip amount, only the gradation skip amount that can be perceived by humans can be extracted.
Furthermore, since the gradation skip amount obtained by subtracting the threshold value equivalent is corrected using the color information, the gradation skip amount actually perceived by humans can be obtained.
Further, since the gradation jump amount obtained in this way is corrected so as to have the same rate as human vision, the gradation skip amount actually perceived by humans can be obtained.

ここで、ＳＬ＊：Ｌ＊の評価値、Ｓａ＊：ａ＊の評価値、Ｓｂ＊：ｂ＊の評価値である。 [Second Embodiment]
The configuration of the image evaluation apparatus 10 in the second embodiment is the same as that of the first embodiment, except that it has a comprehensive evaluation value calculation unit 110 described below.
FIG. 6 is a flowchart showing a processing procedure in the second embodiment. In the second embodiment, input image data is converted into L *, a *, and b * in the L * a * b * color space, and an evaluation value is obtained for each according to the procedure described in the first embodiment. The procedure until obtaining each evaluation value is the same as that in the first embodiment (steps S01 to S05). Then, the comprehensive evaluation value calculating unit 110 calculates the comprehensive evaluation value S2 by weighting each colorimetric value according to the ratio of contribution to human vision (step S06). An example of a formula for calculating the comprehensive evaluation value S2 is shown in Formula 5.

Here, the evaluation value of SL *: L *, the evaluation value of Sa *: a *, and the evaluation value of Sb *: b *.

Equation 5 is an expression when the colorimetric values L *, a * and b * in the L * a * b * color space are used, and the calculation formula of the comprehensive evaluation value is different if the color space is different. Further, the comprehensive evaluation value may be obtained by weighting color information other than the colorimetric value in the uniform color space.
Accordingly, in the evaluation of the color gradation image, it is possible to perform evaluation in consideration of human visual characteristics and further in consideration of the ratio of each colorimetric value to human vision.

[Modification]
The present invention is not limited to the form described above, and can be implemented in various forms. For example, the embodiment described above can be modified as follows.
The image feature amount calculation unit 104 may perform correction using a normal distribution function instead of correcting the brightness distribution with the spatial frequency characteristics of human vision. In this case, a Laplacian of a normal distribution function is used for the lightness distribution. Further, the normal distribution function may be converted into a weighting coefficient matrix and used, or smoothing filter processing may be performed instead of the normal distribution function. A low-pass filter may be used.

When the image feature amount calculation unit 104 calculates the gradation skip amount of the image to be evaluated, it may be calculated from the first derivative of the lightness distribution.
The image to be evaluated is not limited to a gradation image, and may be a uniform image.

  In the first embodiment, when noise generated in the image input unit 101 exists, in order to prevent the noise from being emphasized too much, the brightness distribution averaged in an n × n (n is an integer) region is used. However, the operator may be multiplied after the image size is reduced. For example, since a general reflection original has a high visual sensitivity of about 1 cyc / mm at a clear viewing distance of about 40 cm, the image size may be reduced so that the minimum imaging unit is 1 mm.

1 is a diagram illustrating a configuration of an image evaluation device 10. FIG. 2 is a diagram showing a module group virtually formed in the image evaluation apparatus 10. FIG. It is a figure which shows the example of a Laplacian operator. It is a figure which shows the example of a Laplacian operator. It is a figure which shows the flow of the process in 1st Embodiment. It is a figure which shows the flow of the process in 2nd Embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Image evaluation apparatus, 11 ... CPU, 12 ... RAM, 13 ... Non-volatile memory, 14 ... Communication IF, 20 ... Image input device, 101 ... Image input means, 102 ... Color information conversion means, 103 ... Spatial frequency information conversion Means 104: Image feature amount calculation means 105 ... Evaluation value calculation means 106 ... Threshold value calculation means 107 ... Subtraction means 108 ... First correction means 109 ... Second correction means 110 ... Comprehensive evaluation value calculation means.

Claims (8)

An image input means for reading an image to be evaluated and inputting image data representing a pixel value and a pixel position of the image to be evaluated;
Color information conversion means for converting the pixel value into color information representing any one of lightness, saturation, hue, density, luminance, chromaticity, and intensity;
A spatial frequency information converting means for correcting the color information by a spatial frequency characteristic of human vision;
Image feature amount calculating means for obtaining a difference between adjacent pixels of the color information corrected by the spatial frequency information converting means;
A gradation jump having a gradation jump amount that can be perceived by humans is extracted by comparing the difference obtained by the image feature amount calculation unit with a threshold value of a gradation jump amount that can be perceived by human beings. An image evaluation apparatus comprising: an evaluation value calculating means for obtaining a gradation skip amount that a human actually perceives about a jump. The image evaluation apparatus according to claim 1, further comprising a threshold value calculation unit that obtains a threshold value of a gradation jump amount perceivable by the human using color information. The image evaluation apparatus according to claim 1, further comprising a subtracting unit that subtracts the threshold equivalent amount from the gradation skip amount obtained by the evaluation value calculating unit. The image evaluation apparatus according to claim 3, further comprising: a first correction unit that corrects a gradation skip amount obtained by subtracting a threshold equivalent amount by the subtraction unit using the color information. 5. The image forming apparatus according to claim 4, further comprising: a second correcting unit that corrects the gradation skip amount corrected by the first correcting unit so as to have a rate equal to human perception. An image input means for reading an image to be evaluated and inputting image data representing a pixel value and a pixel position of the image to be evaluated;
Color information conversion means for converting the pixel value into a plurality of types of color information among lightness, saturation, hue, density, luminance, chromaticity, and intensity;
Spatial frequency information converting means for correcting each of the color information by the spatial frequency characteristics of human vision;
Image feature amount calculating means for obtaining a difference between adjacent pixels of the color information corrected by the spatial frequency information converting means;
A gradation jump having a gradation jump amount that can be perceived by humans is extracted by comparing the difference obtained by the image feature amount calculation unit with a threshold value of a gradation jump amount that can be perceived by human beings. An evaluation value calculation means for obtaining a gradation skip amount actually perceived by a human about a jump;
An image evaluation apparatus comprising: comprehensive evaluation value calculation means for performing weighting on each of the gradation skip amounts obtained by the evaluation value calculation means based on a contribution ratio with respect to human perception and adding together. An image input step of reading the image to be evaluated and inputting image data representing a pixel value and a pixel position of the image to be evaluated;
A color information conversion step for converting the pixel value into color information representing any one of lightness, saturation, hue, density, luminance, chromaticity, and intensity;
A spatial frequency information conversion step for correcting the color information by a spatial frequency characteristic of human vision;
An image feature amount calculating step for obtaining a difference between adjacent pixels of the color information corrected by the spatial frequency information converting means;
A gradation jump having a gradation jump amount that can be perceived by humans is extracted by comparing the difference obtained by the image feature amount calculation unit with a threshold value of a gradation jump amount that can be perceived by human beings. An image evaluation method comprising: an evaluation value calculating step for obtaining a gradation skip amount actually perceived by a human with respect to a jump. Computer equipment,
An image input means for reading an image to be evaluated and inputting image data representing a pixel value and a pixel position of the image to be evaluated;
Color information conversion means for converting the pixel value into color information representing any one of lightness, saturation, hue, density, luminance, chromaticity, and intensity;
A spatial frequency information converting means for correcting the color information by a spatial frequency characteristic of human vision;
Image feature amount calculating means for obtaining a difference between adjacent pixels of the color information corrected by the spatial frequency information converting means;
A gradation jump having a gradation jump amount that can be perceived by humans is extracted by comparing the difference obtained by the image feature amount calculation unit with a threshold value of a gradation jump amount that can be perceived by human beings. A program for functioning as an evaluation value calculation means for determining the amount of gradation skip that humans actually perceive about skipping.

Images