JP2006237676A - Method and apparatus for image evaluation, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technology capable of quantitatively and stably evaluating overall image quality sensed by humans with high accuracy as to an image output from a color image output machine in a good correlation with human senses. <P>SOLUTION: Each of evaluation value calculation sections 220, 240, 260 calculates an original image quality element evaluation value Q10 on the basis of original image data D10 and captures input image data D20 corresponding to the original image data D10 to calculate an input image quality element evaluation value Q20. An output image quality evaluation value calculation section 30 compares the original image quality element evaluation value Q10 with the input image quality element evaluation value Q20 to calculate an overall image quality evaluation value Q for denoting the overall image quality of an output image. The comparison not between pixels each but between the evaluation values obtained from the input image and the original image can disregard the effect of the displacement between the input image and the original image, and the image quality element evaluation value can accurately and simply be calculated, resulting in that the evaluation with high accuracy can be attained. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention executes an image evaluation method and image evaluation apparatus for evaluating the image quality of an image output from an image output apparatus (image forming apparatus) such as a color printer, and an image evaluation process using an electronic computer (computer). The present invention relates to an image evaluation program.

  As a method for evaluating the image quality of the image output from the image output device, there are a subjective evaluation for evaluating the image quality by subjective judgment through human vision, and an objective for calculating the image quality feature quantity through a measuring instrument. There is a positive evaluation. Subjective evaluation has a problem of evaluation accuracy such as the evaluation result varies due to human judgment, and in recent years, many objective evaluation methods have been proposed (see, for example, Patent Documents 1 to 4).

JP-A-9-153136 JP 2003-16443 A JP-A-11-39486 JP 2000-251076 A

  For example, in the mechanism described in Patent Document 1, first, image information of an image to be evaluated is converted into spatial frequency distribution information. Then, the spatial frequency distribution information is subjected to filtering correction by a function representing the spatial frequency characteristic of the human visual system according to the observation parameter. Alternatively, when a specific frequency pattern is provided, filtering correction is performed using a function representing the spatial frequency characteristic of the human visual system adapted to the specific frequency pattern. Thereafter, an image evaluation value is calculated from the image information obtained by the inverse transformation. In other words, as a method of evaluating the granularity of a color image, the spatial frequency sensitivity of the human eye according to the viewing conditions for the lightness, saturation, and hue signals of the patch image that was originally output as uniform. Evaluation values such as standard deviation are calculated after filtering according to.

  In the mechanism described in Patent Document 2, a gradation image in which gradation continuously changes is used as an evaluated image as a method for evaluating gradation. In addition, regarding the image to be evaluated, the image quality evaluation is executed by extracting a feature amount from at least one of the brightness information, density information, intensity information, saturation information, chromaticity information, and hue information of the gradation image. Specifically, this image to be evaluated is filtered by the spatial frequency characteristics of the human visual system in the opposite color space, converted to a uniform color space, and then compared with smooth virtual image information to obtain the feature amount. I am trying to calculate.

  Further, in the mechanism described in Patent Document 3, a comprehensive image evaluation value is obtained from the sharpness, granularity, and gradation characteristics of the image to be evaluated, and the image quality is evaluated based on the comprehensive evaluation value. Specifically, based on a linear equation with sharpness, granularity, and gradation characteristics calculated using a pattern image as variables, an evaluation value corresponding to the overall image quality felt by humans is calculated. I am doing so.

  In the mechanism described in Patent Document 4, original image data is stored, an output image (for example, a pattern image) of the original image data by the image output apparatus to be evaluated is directly read by an image input means such as a scanner, and the original image data is read. By associating the pixel of the image data with the pixel of the input image data by the image input means, and comparing the original image data and the input image data for each associated pixel (specifically, taking a difference) The image quality evaluation value corresponding to the subjective evaluation value is calculated by calculating the evaluation value representing the image quality of the output image.

  Here, in the mechanisms described in Patent Documents 1 and 2, the evaluation value of each image quality element is calculated using a pattern image such as a patch image or a gradation image, and the image is evaluated based on the evaluation value itself. . However, various factors contribute to the human sense of total image quality in a complex manner, and cannot be expressed simply by individual image quality degradation factors such as granularity and sharpness. It cannot be said that it represents the quality of the overall image quality that humans feel in the image, and if the evaluation is based on each evaluation value itself, the overall image quality of the image cannot be evaluated with good correlation with human senses. . Hereinafter, this problem is also referred to as a first problem.

  On the other hand, in the mechanism described in Patent Document 3, an overall image evaluation value is obtained from the sharpness, granularity, and gradation characteristics of the image to be evaluated, which are psychophysical quantities, and the image quality is evaluated based on the overall evaluation value. Therefore, it is possible to evaluate the image quality with a good correlation with human senses and to perform a quantitative overall image quality evaluation, thereby solving the first problem.

  However, in the mechanism described in Patent Document 3, when the output image path is different between the pattern image for objective evaluation and the pattern image for subjective evaluation, the image quality element evaluation value of the pattern image and the total image quality evaluation value of the pattern image Is difficult to associate. For example, when the automatic image quality adjustment function of the printer driver is used, color conversion processing depending on the output image is performed, so the color conversion processing at the time of output differs between the pattern image and the pattern image, and the image quality of both images is naturally Come different. As a method for solving this problem, for example, it is conceivable to directly input a pattern.

  However, with this coping method, even if the problem caused by the different output image paths can be solved, the problem caused by the characteristics of the image input apparatus cannot be solved. For example, even if a certain color (Lab, LCH) is input, various RGB values of input image data are obtained depending on the input device. Hereinafter, this problem is also referred to as a second problem.

  Further, in the mechanism described in Patent Document 4, since it is essential to take a difference between pixels for the input image and the original image in calculating the evaluation value, there is a problem of a decrease in accuracy due to the positional deviation between both images. Have. That is, there is often a positional shift between the input image and the original image, and it is difficult to obtain an accurate difference for each pixel, and there is a problem that an accurate image quality evaluation value cannot be calculated. Hereinafter, this problem is also referred to as a third problem.

  Further, in the mechanism described in Patent Document 4, evaluation values of image quality elements such as color reproducibility and graininess are calculated, and images are evaluated based on the evaluation values themselves. Similar to the mechanism described in No. 2, it has the first problem.

  This invention is made | formed in view of the said situation, and it aims at providing the mechanism which can solve at least 1 of the 1st-3rd problem which the said patent documents 1-4 have. In particular, it is an object to provide a mechanism that can solve at least one of the second and third problems, and more preferably, an image quality evaluation value corresponding to a subjective overall image quality is obtained, The objective is to provide a mechanism that can solve the first problem by evaluating the overall image quality with a good correlation with human senses, and realize a more stable and higher evaluation.

  In the image evaluation method according to the present invention, first, an original image quality element evaluation value representing the image quality of the output image is calculated based on the original image data representing the output image. Also, input image data corresponding to the original image data is captured, and an input image quality factor evaluation value representing the image quality of the input image corresponding to the output image is calculated based on the captured input image data. Then, the original image quality factor evaluation value calculated in this way and the input image quality factor evaluation value are compared (for example, a difference is taken) to calculate an output image quality evaluation value representing the image quality of the output image.

  An image evaluation apparatus according to the present invention is an image evaluation apparatus for evaluating the image quality of an image to be evaluated suitable for implementing the image evaluation method according to the present invention, and is based on original image data representing an output image. An original image quality element evaluation value calculation unit for calculating an original image quality element evaluation value representing the image quality of the output image, an input interface unit for acquiring input image data corresponding to the original image data, and an input acquired by the input interface unit An input image quality element evaluation value calculation unit that calculates an input image quality element evaluation value representing the image quality of the input image corresponding to the output image based on the image data, and an original image calculated by the original image quality element evaluation value calculation unit By comparing the image quality factor evaluation value with the input image quality factor evaluation value calculated by the input image quality factor evaluation value calculation unit, the output quality representing the output image quality is expressed. It was assumed and an output image quality evaluation value calculation unit for calculating an image quality evaluation value.

  Furthermore, the program according to the present invention is suitable for realizing the image evaluation method and the image evaluation apparatus according to the present invention by software using an electronic computer. The program may be provided by being stored in a computer-readable storage medium, or may be provided by distribution via wired or wireless communication means.

  Further, the invention described in the dependent claims defines a further advantageous specific example of the image evaluation apparatus, the image evaluation method, or the program according to the present invention.

  For example, when calculating the evaluation value, the influence of the image input apparatus may be eliminated by using a color signal in a device-independent color space. In addition, an evaluation value corresponding to the subjective image quality may be calculated by considering an evaluation result obtained in advance from a subjective evaluation experiment.

  In the present invention, the original image quality element evaluation value of the output image is calculated, the input image quality element evaluation value representing the image quality of the input image corresponding to the output image is calculated, and each of the images directly calculated from both images is calculated. The image quality of the output image is evaluated by comparing the image quality element evaluation values.

  Thereby, the comparison between the evaluation values obtained for the respective images is performed instead of performing the comparison for each pixel (for example, taking a difference), so that the influence of the positional deviation between the input image and the original image can be ignored. The image quality element evaluation value can be calculated accurately and easily. As a result, accurate evaluation can be performed.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

<System configuration>
FIG. 1 is a block diagram illustrating a schematic configuration of an image evaluation system 1 including an image evaluation apparatus. Here, a block diagram according to functional blocks of the processing procedure is shown.

  In the algorithm of the image evaluation method of this embodiment, a case will be described in which the image quality of a color image output from an image output device such as a color printer is evaluated.

  As shown in the figure, the image evaluation system 1 has at least one evaluation value Q10 (for an original image; hereinafter referred to as an original image quality element evaluation value) for the graininess, sharpness, gradation characteristics, and color reproducibility of the image to be evaluated. Q10), Q20 (for the input image; hereinafter referred to as input image image quality element evaluation value Q20), an overall image quality evaluation value Q is obtained, and the image quality is comprehensively determined based on the obtained overall image quality evaluation value Q. An image evaluation apparatus 10 to be evaluated, an image output apparatus (image forming apparatus) 50 that generates (forms) an image to be evaluated on a predetermined recording medium, and an image to be evaluated that is generated by the image output apparatus 50 is an electronic image. And an image input device 70 for inputting the converted input image data D20 to the image evaluation device 10.

  For example, a color printer using an image forming process such as an ink jet method or an electrophotographic method is used as the image output device 50 that generates an image to be evaluated Image to be evaluated. For example, the original image data D10 stored in the image evaluation apparatus 10 is converted into a color signal in a device-dependent color space such as RGB or CMYK to form an image on an output medium such as paper.

  The image input device 70 only needs to be able to generate input image data D20 (digitized image information) corresponding to the original image data D10 and input it to the image evaluation device 10, such as a flatbed scanner. A device that digitizes and outputs pixel information of the image to be evaluated Image generated by the image output device 50 is used. For example, the output image (image to be evaluated Image) of the original image data D10 generated by the image output device 50 is optically read, and the read input image data D20 is supplied to the image evaluation device 10.

  In this case, it is desirable that the image input device 70 has been subjected to characteristic calibration, driver adjustment, and the like, and the graininess, sharpness, gradation characteristics, and color reproducibility do not change depending on the subject. It is further desirable that automatic processing such as AGC and edge enhancement performed on the image input device 70 side is turned off.

  Alternatively, it is not necessary to have an input means for optically reading the output image of the original image data D10, and image data downloaded via the network 9 using a network device connected to an external device or a storage recording medium. (Original image data D10) may be acquired and supplied to the image evaluation apparatus 10 as input image data D20. Alternatively, image data (original image data D10) stored in a portable storage medium such as a CD-ROM may be captured and supplied to the image evaluation apparatus 10 as input image data D20.

  Although not shown, the image evaluation device 10 includes a control unit that controls each functional unit in the device, and a display unit that notifies the user of guidance information, predetermined information processing results, management information, and the like as image information. In addition, a voice notification unit comprising a built-in or externally installed piezoelectric body or speaker that notifies these as voice information, and an instruction input having a mouse or keyboard for receiving various instruction inputs from the operator to the device A user interface unit including a unit (operation unit).

  The image evaluation apparatus 10 includes an output interface unit 11out and an input interface unit 11in as interface function units for delivering image data, and an image output device 50 to be evaluated as functional units related to execution of the image evaluation process. The original image storage unit 12 that stores the original image data D10 necessary for the evaluation of the input image and the input image storage unit 14 that stores the input image data D20 supplied from the image input device 70 are provided.

  The output interface unit 11out outputs the original image data D10 stored in the original image storage unit 12 to various devices outside the apparatus. The input interface unit 11in takes in the input image data D20 from various devices outside the apparatus and passes it to the input image storage unit 14.

As the original image data D10, not only a pattern image such as a patch image or a gradation image but also a pattern image such as a photographic image can be used. In this embodiment, the case of a pattern image will be described. The original image data D10 is preferably device-independent data. As an example, it is preferable to use a color signal in a device-independent color space such as sRGB (standard RGB), XYZ, Lab (correctly L ^* a ^* b ^* ), LCh (correctly L ^* C ^* h ^° ), or the like. .

  Here, the sRGB color space data is a color space defined by the International Electrotechnical Commission (IEC), which is widely used in the fields of peripheral devices for personal computers such as digital cameras, printers, and monitors. The color data conforms to the international standard. By performing color adjustment according to sRGB, the difference in color between input and output can be minimized.

  Further, the image evaluation apparatus 10 includes a color signal conversion unit 16 that converts image data in a device-dependent color space into image data in a device-independent color space. Further, the image evaluation device 10 calculates an original image quality element evaluation value Q10 based on the original image data D10, and inputs image quality element based on the input image data D20 (image data) supplied from the image input device 70. An image quality element evaluation value calculation unit 20 that calculates an evaluation value Q20 is provided. Further, the image evaluation apparatus 10 calculates an output image quality evaluation value for calculating an image quality evaluation value of the output image based on the original image quality element evaluation value Q10 and the input image quality element evaluation value Q20 obtained by the image quality element evaluation value calculation unit 20. A value calculation unit 30 is provided.

  The color signal conversion unit 16 converts the original image data D10 and the input image data D20 into color data in a uniform perceptual color space, a tristimulus value XYZ, or tristimulus value XYZ, which is a device-independent color space. (Collectively, original image data D12, input image data D22). The image quality element evaluation value calculation unit 20 may calculate the evaluation value using the conversion result (original image data D12, input image data D22) of the color signal conversion unit 16.

  In the uniform perceptual color space, since the color difference has a linear relationship with the color difference when viewed with the human eye, it is possible to calculate an evaluation value that is intuitively understandable. The tristimulus values XYZ defined by CIE represent absolute color values that do not depend on device characteristics such as an image input device. In this aspect, the original image data and the input image data are created in the XYZ color system that represents an absolute color or a color system that can be converted into the color system. Even if the input color characteristic is not considered at all, it can be compared as color information that matches the visual sensitivity of the human eye, and an evaluation value corresponding to the subjective evaluation value can be calculated.

  The image quality element evaluation value calculation unit 20 is a granularity evaluation value calculation unit 220 that calculates an evaluation value related to graininess (granularity evaluation value Qg) for each of the original image and the input image, and an evaluation value related to sharpness (sharpness). A sharpness evaluation value calculation unit 240 that calculates an evaluation value Qs), an evaluation value related to gradation characteristics (tone evaluation value Qt), and an evaluation value related to color reproducibility (color reproducibility evaluation value Qc). And a gradation / color reproducibility evaluation value calculation unit 260.

<Calculation of individual evaluation values>
Based on the color signals D12 and D22 of the device-independent color space subjected to color conversion by the color signal conversion unit 16, the image quality element evaluation value calculation unit 20 is an evaluation value of the image quality element for each of the original image and the input image. A certain original image quality factor evaluation value Q10 and an input image quality factor evaluation value Q20 are calculated.

  The image quality element evaluation value calculation unit 20 calculates an evaluation value related to at least one of graininess, sharpness, gradation, and color reproducibility as the evaluation value of the image quality element. Preferably, the plurality of evaluation values are calculated so as to contribute to the comprehensive evaluation. More preferably, in order to improve the accuracy of the overall image quality evaluation, evaluation values relating to granularity, sharpness, gradation characteristics, and color reproducibility having high correlation with psychological evaluation for human image quality are calculated. .

  In addition, the image quality element evaluation value calculation unit 20 preferably uses device-independent pixel information as an evaluation value to solve the problem in the case where the output image path differs between the objective evaluation pattern image and the subjective evaluation pattern image. It is desirable to use for calculation.

  Further, the image quality element evaluation value calculation unit 20 preferably calculates an evaluation value directly from each image first, and then solves the problem caused by taking a difference between the pixels of the input image and the original image. It is desirable to take a difference between the evaluation values of each image.

  Further, it is desirable to calculate a highly accurate evaluation value by decomposing the color information into color attributes such as brightness difference, saturation difference, hue difference, or color difference and performing evaluation for each attribute. For example, each of the evaluation value calculation units 220, 240, and 260 directly compares the brightness difference from the target image for each of the original image data D10 (preferably the original image data D12) and the input image data D20 (preferably the input image data D22). More preferably, (ΔL *), saturation difference (ΔC *), hue difference (ΔH °), or color difference (ΔE) is calculated, and the respective evaluation values are calculated based on these components.

  In this case, the evaluation value is calculated by the mean square of the color difference signal of each pixel, or the spatial frequency analysis is performed on the color difference signal, and the spatial frequency analysis result corresponds to the sensitivity distribution with respect to the spatial frequency of the human eye. The evaluation value may be calculated by superimposing the weighting functions.

  By using the color data of the uniform perceptual color space converted by the color signal conversion unit 16 and the device independent color signals (color data; D12, D22) that can be converted into tristimulus values XYZ or tristimulus values XYZ, The influence of the input color characteristics of the image input device can be eliminated. For example, even if a certain color (Lab, LCH) is input, RGB values of various input image data can be obtained depending on the image input device. However, an input characteristic profile (color conversion profile) of the image input device is used. By converting to device-independent data, the same Lab and LCh can be obtained even if the image input devices are different.

  In order to solve the problem caused by the different output image paths in the evaluation of the pattern image, a method of directly inputting the pattern is adopted. When the automatic image quality adjustment function is applied, the output color conversion varies depending on the image. Therefore, the output image path differs between the pattern image and the pattern image, and the input image data (RGB) also differs. Even if the color signal is converted to a device-independent color signal, the values of Lab and LCh are naturally different between the two images. For this reason, since the object to be evaluated is only a pattern image, the pattern image is directly input.

  Here, as a method for calculating the evaluation value regarding the graininess, for example, after filtering the lightness, saturation, and hue signal of the patch image output so as to be originally uniform based on the spatial frequency characteristics of the human eye There is a first method for calculating the evaluation value of the standard deviation. However, since this method uses a pattern image, the granularity of a pattern image such as a photograph cannot be evaluated.

  Further, the spatial frequency component is calculated using at least one optical information among the density information, lightness information, or chromaticity information of the image to be evaluated, and the calculation result is corrected according to the spatial frequency characteristics of the visual system. There is also a second method for calculating an evaluation value related to graininess by performing a correction according to the average characteristic of the optical information on a value obtained by integrating the correction value. In this method, it is possible to calculate a highly accurate evaluation value having a good correlation with human senses. In addition, by adding correction according to the average characteristic of the optical information of the image to the integration result, it is possible to calculate an evaluation value related to the graininess independent of the characteristic of the optical information of the image.

  Alternatively, two-dimensional orthogonal transformation is performed on the evaluation target image to generate information indicating a two-dimensional spatial frequency distribution, which is converted into one dimension, and then corrected by a function representing the spatial frequency characteristics of the human visual system. There is also a third method for calculating an evaluation value related to graininess by integration.

  In the second and third methods, since the evaluation value related to the graininess is calculated without using the pattern image, the graininess of the pattern image can also be evaluated.

  In any case, it is preferable to evaluate the image of the color signal in the device-independent color space. For example, it is preferable to perform two-dimensional orthogonal transformation on the image of the color signal in the device-independent color space. By obtaining an evaluation value related to graininess from a color signal in a device-independent color space, it is possible to improve a decrease in evaluation accuracy related to graininess caused by a difference in image input paths.

  As a method for calculating an evaluation value related to sharpness, for example, a first method is used in which a black and white line repeating pattern is output, and a contrast is calculated from the density profile to obtain a sharpness (resolution) evaluation value. There is. However, since this method uses a pattern image, the sharpness of a pattern image such as a photograph cannot be evaluated.

  Alternatively, a spatial frequency component (spatial frequency distribution information) is calculated using at least one optical information of density information, lightness information, or chromaticity information of the image to be evaluated, and an observation parameter (for example, an observation distance) is calculated. Obtained by inverse conversion from the correction value by performing a filtering correction by a function representing a spatial frequency characteristic of the visual system in accordance with an observation parameter including at least one of the luminance of the evaluated image and the ambient luminance of the evaluated image) There is also a second method in which an evaluation value related to sharpness is calculated by calculating an evaluation value related to sharpness from image information, or by performing correction according to an average characteristic of optical information on a value obtained by integrating this correction value. In this method, it is possible to calculate a highly accurate evaluation value having a good correlation with human senses. In addition, by adding correction according to the average characteristic of the optical information of the image to the integration result, it is possible to calculate an evaluation value related to sharpness that does not depend on the characteristic of the optical information of the image.

  Alternatively, there is a third method for calculating the edge value of the image in the horizontal direction and the vertical direction with respect to the evaluation target image by the differential filter, and calculating the evaluation value relating to the sharpness from the average of the square root of the square sum of the edge amount. is there.

  In the second and third methods, since the evaluation value related to the sharpness is calculated without using the pattern image, the sharpness of the pattern image can also be evaluated.

  In any case, it is preferable to evaluate the image of the color signal in the device-independent color space. For example, it is preferable to perform differential filter processing on an image of a color signal in a device-independent color space. By obtaining an evaluation value related to sharpness from a color signal in a device-independent color space, it is possible to improve a decrease in evaluation accuracy related to sharpness caused by a difference in image input paths.

  Further, as a method for calculating an evaluation value related to gradation characteristics, there is a first method in which a gradation image whose gradation changes continuously is used as an evaluation image. At this time, a feature amount is extracted from at least one of lightness information, density information, intensity information, saturation information, chromaticity information, and hue information of the gradation image as the evaluation image to obtain an evaluation value.

  There is also a second method for measuring a perceived brightness difference or density difference with respect to images output at adjacent gradation levels. For example, a difference in optical information between adjacent gradation levels is calculated using at least one optical information among density information, lightness information, or chromaticity information of the image to be evaluated, and a human result for the optical information is calculated as the calculation result. Correction according to the recognition limit is performed, and this correction value is integrated at all gradation levels. In this method, it is possible to calculate a highly accurate evaluation value having a good correlation with human senses. In addition, by correcting the difference in optical information according to human perception limits on optical information, it is possible to calculate highly accurate image evaluation values that correlate well with human senses and improve overall image quality evaluation accuracy Can be made.

  In general, in a printer or a digital image processing system in which input / output gradations are discrete, a third method in which the relative numerical relationship of the output level with respect to the input level and the input / output curve (γ curve) is used as the gradation characteristic. There is.

  Alternatively, there is a fourth method of outputting a gradation image in which gradation changes continuously and measuring the change in brightness, saturation, or hue to calculate an evaluation value of gradation that is felt by humans.

  In the first to fourth methods, since the pattern image is used, it is not possible to evaluate the gradation of a picture image such as a photograph.

  Further, there is a fifth method for calculating the brightness difference (ΔL *) from the difference for each pixel of the image and using the input image and the original image as evaluation values of the gradation characteristics. However, in this method, it is necessary to obtain a difference for each pixel between the input image and the original image, and it is not possible to evaluate the gradation characteristics of the image alone.

  Alternatively, there is a sixth method for calculating the number density for an arbitrary color region with respect to the evaluation target image and calculating the evaluation value from the standard deviation of the frequency distribution of the number density.

  In the fifth and sixth methods, since the evaluation value related to the gradation characteristic is calculated without using the pattern image, the gradation characteristic of the pattern image can also be evaluated. In the sixth method, since the evaluation value related to the gradation characteristics is calculated from the evaluation target image itself, the evaluation value related to the gradation characteristics can be directly calculated from one evaluation image without the reference image. .

  In any case, it is preferable to evaluate the image of the color signal in the device-independent color space. By obtaining the evaluation value related to the gradation characteristic from the color signal in the device-independent color space, it is possible to improve the decrease in evaluation accuracy regarding the gradation characteristic caused by the difference in the image input path.

  As an evaluation value calculation method for color reproducibility, brightness difference (ΔL *), saturation difference (ΔC *), and hue difference (ΔH °) from the difference of each pixel of the input image and the original image. There is a first method of calculating a color difference (ΔE) and using it as an evaluation value of color reproducibility. However, in this method, it is necessary to obtain a difference for each pixel between the input image and the original image, and the color reproducibility of the image alone cannot be evaluated.

  Alternatively, there is a second method for calculating the number density for an arbitrary color region with respect to the evaluation target image and calculating the evaluation value from the average value of the frequency distribution of the number density. In this method, since the evaluation value related to color reproducibility is calculated from the evaluation target image itself, the evaluation value related to color reproducibility can be directly calculated from one evaluation image without the reference image.

  Further, it is preferable to evaluate the color signal image in the device-independent color space. By obtaining an evaluation value related to color reproducibility from a color signal in a device-independent color space, it is possible to improve a decrease in evaluation accuracy related to color reproducibility due to a difference in image input paths.

  In addition to the examples given above, various methods have been proposed for obtaining evaluation values for graininess, sharpness, gradation characteristics, and color reproducibility, and they may be used. As a measurement physical quantity for obtaining graininess, sharpness, gradation characteristics, and color reproducibility, various methods have been proposed depending on whether density or brightness is used in addition to the examples given above. You may use them.

<Calculation of comprehensive evaluation value>
The output image quality evaluation value calculation unit 30 determines the granularity evaluation value Qg related to the measured granularity, the sharpness evaluation value Qs related to the sharpness, the gradation evaluation value Qt related to the gradation characteristics, and the color reproducibility. Using the related color reproducibility evaluation value Qc, an overall image quality evaluation value Q that represents the overall image quality felt by humans with a good correlation with human senses is obtained.

  By the cooperative processing of the color signal conversion unit 16, the image quality element evaluation value calculation unit 20, and the output image image quality evaluation value calculation unit 30, an image to be evaluated and a measurement image (under the same conditions as the evaluation target image) Input image) to calculate at least one of graininess, sharpness, gradation, and color reproducibility, so that the overall image quality of the image correlates well with the human sense without depending on the device. To be able to evaluate. Also, when calculating the evaluation value, the total feature quantity Q corresponding to the subjective image quality is calculated by considering the evaluation result obtained from the subjective evaluation experiment in advance.

  For example, based on a plurality of evaluation values related to graininess, sharpness, gradation characteristics, and color reproducibility, the product of graininess, sharpness, and gradation characteristics that visually represents the “roughness” of the image, Further, it is also possible to obtain a higher-accuracy comprehensive image evaluation of the image by a value calculated from a linear equation with the reciprocal of sharpness representing the “blur” of the image as a variable. In this case, the accuracy of the overall image quality evaluation is improved by using a value obtained by correcting the product of graininess, sharpness and gradation characteristics, and the reciprocal value of sharpness by a power of a predetermined constant. It is good to.

  By quantitatively obtaining an overall image evaluation value (overall image quality evaluation value Q) of an image when two image deterioration factors such as “roughness” and “blur” of the image contribute in a complicated manner, It becomes possible to comprehensively evaluate sharpness. In addition, by correcting from a power of a predetermined constant, it is possible to accurately capture the influence on the human sense of the image when the graininess, sharpness, or gradation characteristics are slightly changed.

  Further, the accuracy of the overall image quality evaluation may be improved by using a value obtained by correcting the product of graininess, sharpness and gradation characteristics, and the reciprocal value of sharpness by a power of a predetermined constant. Good. Alternatively, by using a value suitable for each image type as a coefficient of a predetermined linear equation, a comprehensive image quality evaluation corresponding to the characteristics of the image deterioration factor of each image type may be performed.

  Alternatively, the overall image quality evaluation value Q may be calculated using a weighted linear regression equation (for example, weighted linear sum) of the image quality element evaluation values Q10 and Q20. For example, for each of the lightness difference, saturation difference, hue difference, or color difference, a mean square value is obtained, and an evaluation value is calculated based on the mean square result, or the lightness difference, saturation difference, hue difference is calculated. Alternatively, spatial frequency analysis is performed for each color difference, and a weight function corresponding to the spatial frequency sensitivity distribution of the human eye is superimposed on each spatial frequency analysis result, and an evaluation value is calculated based on the superimposition result. Good. Since the difference between the evaluation values can be combined into one value by taking the square-square average, the comprehensive image quality evaluation value Q that is easy to handle can be calculated.

  Alternatively, a square-square average is obtained for each of the lightness difference, saturation difference, hue difference, or color difference, and the total image quality evaluation value Q is calculated based on a plurality (preferably all four) of these square-square average results. May be calculated. In addition, spatial frequency analysis is performed for each of the lightness difference, saturation difference, hue difference, or color difference, and a weight function corresponding to the spatial frequency sensitivity distribution of the human eye is superimposed on each spatial frequency analysis result. The total image quality evaluation value Q may be calculated based on the result. In this method, the spatial frequency sensitivity of the human eye can be considered for each attribute of the color information, and an overall image quality evaluation value Q having a higher correlation with the subjective evaluation value can be calculated.

  Rather than simply evaluating the image based on individual image quality degradation factors such as graininess and sharpness, the image quality degradation factor of the image to be evaluated is the psychophysical quantity of graininess, sharpness, gradation characteristics, And comprehensive evaluation of image quality by obtaining a comprehensive evaluation index value (total image quality evaluation value Q) by combining multiple evaluation values of color reproducibility Can be carried out quantitatively and stably. In addition, by using the combined amount of psychophysical quantities as a variable, it is possible to calculate a highly accurate image evaluation value that has a good correlation with human senses.

  The apparatus for reading the measurement image for evaluating the graininess, sharpness, gradation characteristics, and color reproducibility is not limited to a flatbed scanner, but may be, for example, a microdensitometer, a drum scanner, A colorimeter, a densitometer, etc. can also be used.

  In addition to the examples given above, various methods have been proposed for the spatial frequency component derivation method, the correction method based on human visual characteristics, and the correction method for the average characteristic of the image. Good.

  In any case, the image evaluation apparatus 10 of the present embodiment uses the respective evaluation values related to the obtained graininess, sharpness, gradation characteristics, and color reproducibility to obtain a comprehensive image quality felt by humans. It is only necessary to be able to calculate the overall image quality evaluation value Q that is well correlated with the human senses. Preferably, the evaluation value is directly calculated from one image so that accurate image (particularly image quality) evaluation can be performed even when there is no reference image. Further, when the image input devices are different, the color signal converted into a device-independent color signal by the color signal conversion unit 16 is used for calculation of the evaluation value so that accurate image (particularly image quality) evaluation can be performed. Hereinafter, preferred embodiments will be specifically described.

<Configuration of color signal converter>
FIG. 2 is a block diagram illustrating a schematic configuration of the color signal conversion unit 16. Here, a block diagram according to functional blocks of the processing procedure is shown.

  As shown in the figure, the color signal conversion unit 16 converts the original image data D10 stored in the original image storage unit 12 into color data in a uniform perceptual color space, tristimulus value XYZ, or tristimulus value XYZ. The original image data color conversion unit 162 for converting the device-independent color signal and the input image data D20 stored in the input image storage unit 14 into color data, tristimulus values XYZ or tristimulus values in a uniform perceptual color space And an input image data color conversion unit 164 for converting color data that can be converted into XYZ into device-independent color signals.

  For example, the original image data color conversion unit 162 temporarily converts the original image data D10 indicated by the sRGB values captured from the original image storage unit 12 into XYZ tristimulus values that are color signals in a device-independent color space. Later, the image data is converted into original image data D12 represented by Lab values that are uniform color spaces or L (lightness) C (saturation) h (hue) values constituting three attributes of color.

  Further, as an example, the input image data color conversion unit 164 temporarily converts the input image data D20 indicated by the RGB values captured from the input image storage unit 14 into XYZ tristimulus values that are color signals in a device-independent color space. After that, the image data is converted into input image data D22 indicated by a Lab value that is a uniform color space or L (lightness) C (saturation) h (hue) value that constitutes three attributes of color.

  Here, since the RGB value of the input image data D20 is a device-dependent color signal, the input image data color conversion unit 164 uses the input characteristic profile (color conversion profile) of the image input device 70, and is device-independent. The color signal is converted into XYZ3 stimulus values or Lab and LCh.

  Further, the input image data color conversion unit 164 filters the input image data D20 not only with respect to color and gradation characteristics but also with respect to the spatial frequency (MTF) characteristics in order to correct the input characteristics of the image input device 70. Make corrections. After this correction, it is converted into an XYZ3 stimulus value or Lab value or LCh value which is a device independent color signal. That is, in the configuration of the present embodiment, the input image data color converting unit 164 generates input image data (for example, input image data) before generating the input image quality factor evaluation value from the input image. It functions as an input characteristic correction unit that corrects the input characteristics of the image input device 70) that generates the data D20.

<Configuration of graininess evaluation value calculation unit>
FIG. 3 is a block diagram illustrating a schematic configuration of the granularity evaluation value calculation unit 220 of the image quality element evaluation value calculation unit 20. Here, a block diagram according to functional blocks of the processing procedure is shown.

  The granularity evaluation value calculation unit 220 is obtained by the color signal conversion unit 16 in calculating the granularity evaluation value Qg (Qg10, Qg20) for each of the original image and the input image. Lightness information L (x, y), saturation information C (x, y), and hue information h (x, y) are used. Then, the granularity evaluation value Qg representing the degree of image noise (roughness) is subjected to two-dimensional orthogonal transformation on the image of the color signal in the device-independent color space, and information indicating a two-dimensional spatial frequency distribution is obtained. After being generated and made one-dimensional, it is corrected by a function representing the spatial frequency characteristic of the human visual system and is calculated by integration.

  Specifically, as shown in the figure, the graininess evaluation value calculation unit 220 firstly has the lightness information L (x, y), saturation information C (x, y), hue obtained by the color signal conversion unit 16. Color difference calculation unit 222 for obtaining color difference information based on information h (x, y), and information indicating a two-dimensional spatial frequency distribution by performing two-dimensional orthogonal transformation on the color difference information obtained by color difference calculation unit 222 And a two-dimensional orthogonal transformation unit 224 for generating

  The granularity evaluation value calculation unit 220 is a two-dimensional power spectrum calculation unit that calculates a two-dimensional power spectrum representing the intensity of the two-dimensional spatial frequency F (u, v) that has been two-dimensionally orthogonally transformed by the two-dimensional orthogonal transformation unit 224. 226 and a two-dimensional power spectrum one-dimensional calculation unit 228 that integrates the power spectrum for the same spatial frequency obtained by the two-dimensional power spectrum calculation unit 226 and calculates a one-dimensional power spectrum.

  In addition, the granularity evaluation value calculation unit 220 multiplies the power spectrum that has been made one-dimensional by the two-dimensional power spectrum one-dimensionalization calculation unit 228 with a function that represents the spatial frequency characteristics of the visual system, and integrates it. A visual spatial frequency characteristic correction unit 230 that calculates information serving as an index of image noise, and two-dimensional lightness, two-dimensional saturation, and two-dimensional hue based on the index information obtained by the visual spatial frequency characteristic correction unit 230 And a granularity evaluation value calculation unit 232 that calculates a granularity evaluation value Qg that is an index of image noise.

  Here, the color difference calculation unit 222 performs lightness difference information ΔL (x, x, y) that is a difference between the lightness information L (x, y) and the lightness average value Lave according to the following formulas (1-1) and (1-2). y) and saturation information ΔC (x, y), which is the difference between the saturation information C (x, y) and the saturation average value Cave, is calculated.

  Further, the color difference calculation unit 222 calculates hue difference information Δh (x, y) that is a difference between the hue information h (x, y) and the hue average value have according to the following formula (2-1). However, the hue difference information ΔH (x, y) in a format converted from an angle to a distance is calculated according to the following equation (2-2). “^” Indicates a power.

  The two-dimensional orthogonal transform unit 224 performs a discrete two-dimensional Fourier transform for each image divided into lightness difference information ΔL (x, y), saturation difference information ΔC (x, y), and hue difference information ΔH (x, y). To obtain the spatial frequency component. Specifically, in accordance with the following formulas (3-1), (3-2), and (3-3), information ΔL (x, y), ΔC (x, y), ΔH (x , Y) Information FΔL (u, v), FΔC (u, v), FΔH (u, v) after the two-dimensional Fourier transform is calculated.

  The two-dimensional power spectrum calculation unit 226 performs Fourier transform information FΔL (u, u, obtained by the two-dimensional orthogonal transform unit 224 according to the following formulas (4-1), (4-2), and (4-3). v), FΔC (u, v), FΔH (u, v), two-dimensional power spectra PΔL (u, v), PΔC (u, v) representing the intensity of the two-dimensional spatial frequency F (u, v), PΔH (u, v) is calculated.

  The two-dimensional power spectrum one-dimensionalization calculation unit 228 integrates the power spectrum for the same spatial frequency to calculate one-dimensional PΔL (f), PΔC (f), and PΔH (f) in order to simplify the measurement.

  The luminosity spatial frequency characteristic correction unit 230 performs power spectrum PΔL one-dimensionalized by the two-dimensional power spectrum one-dimensional operation unit 228 according to the following formulas (5-1), (5-2), and (5-3). (F), PΔC (f), and PΔH (f) are multiplied by functions VTFL (f), VTFC (f), and VTFH (f) representing the spatial frequency characteristics of the visual system and integrated in the frequency domain. To calculate QΔL, QΔC, and QΔH, which are indicators of image noise.

  As the functions VTFL (f), VTFC (f), and VTFH (f), for example, the following equation (6) is used.

  In addition, you may add correction | amendment according to the average brightness (average characteristic of optical information) of an image as needed.

  The granularity evaluation value calculation unit 232 adds an appropriate weight to each according to the following formula (7), and adds an index of image noise combining two-dimensional brightness, two-dimensional saturation, and two-dimensional hue. The graininess evaluation value Qg is calculated. Here, the smaller the granularity evaluation value Qg, the smaller the image noise and the better the granularity.

  In the formula, a, b, and c are weights for lightness, saturation, and hue. These weights depend on the image reading device used, the image reading method, the measured physical quantity or optical information, the color system, the visual characteristic correction function or correction method, and various correction formulas or correction methods for the average characteristic of the image. Set the optimal value.

  Note that the above calculation process for calculating the evaluation value (granularity evaluation value Qg) related to the graininess is merely an example, and various changes can be made. For example, in the above example, Fourier transform is used when obtaining a spatial frequency component, but wavelet transform may be used. In the above example, the two-dimensional power spectrum one-dimensionalization calculation unit 228 converts the two-dimensional spatial frequency component into one dimension for simplification of the calculation, but the conversion to one dimension may not be performed. Further, the correction of the human visual frequency characteristic may be performed after the two-dimensional spatial frequency component is converted into one dimension, or may be performed before the conversion.

<Configuration of sharpness evaluation value calculation unit>
FIG. 4 is a block diagram illustrating a schematic configuration of the sharpness evaluation value calculation unit 240 of the image quality element evaluation value calculation unit 20. Here, a block diagram according to functional blocks of the processing procedure is shown.

  The sharpness evaluation value calculation unit 240 calculates brightness information L (x, y) obtained by the color signal conversion unit 16 in calculating the sharpness evaluation value Qs (Qs10, Qs20) for each of the original image and the input image. Use. In addition, the sharpness evaluation value calculation unit 240 is generally called MTF (Modulation Transfer Function) for a color signal image in a device-independent color space, and is a spatial frequency component of a component representing edge sharpness characteristics such as a line. Are obtained for each of the horizontal and vertical directions of the image, and the sharpness evaluation value Qs is calculated based on them.

  For example, the edge amount of the image in the horizontal direction and the vertical direction is calculated by a differential filter, and the sharpness evaluation value Qs is calculated from the average of the square root of the square sum of the edge amount. Specifically, as illustrated, the sharpness evaluation value calculation unit 240 includes an edge amount calculation unit 242 that calculates the horizontal and vertical edge amounts of the lightness information L (x, y), and an edge amount calculation. An edge strength calculation unit 244 that calculates edge strength for each pixel based on the edge amount obtained by the unit 242 and an image sharpness index based on the edge strength obtained by the edge strength calculation unit 244 A sharpness evaluation value calculation unit 246 for obtaining a sharpness evaluation value Qs.

  For example, the edge amount calculation unit 242 uses a Sobel filter (differential filter) represented by the following formulas (8-1) and (8-2), and uses the edge amount SLh (in the horizontal direction) of the brightness information L (x, y). x, y) and the vertical edge amount SLv (x, y) are calculated. The calculation of the edge amount (edge image) is not limited to such a method, and for example, a method of creating an image including only a high-frequency component using unsharp mask processing can be used.

  The edge strength calculation unit 244 calculates the edge strength SL (x, y) for each pixel according to the following formula (9).

  Further, the sharpness evaluation value calculation unit 246 extracts pixels whose edge strength obtained by the edge strength calculation unit 244 is larger than a threshold value (for example, “0”), and averages the edge strength according to the following equation (10). A value SLm is calculated, and this is defined as a sharpness evaluation value Qs that is an index of sharpness. Here, the larger the sharpness evaluation value Qs, the higher the sharpness of the image. Note that the sharpness evaluation value Qs is not limited to a simple average value, and may be a root mean square (RMS) or a maximum value.

  Note that the above calculation process for calculating the sharpness evaluation value (sharpness evaluation value Qs) is merely an example, and various changes can be made. For example, the sharpness evaluation value calculation unit 240 obtains the spatial frequency component of the component representing the sharpness characteristic of the edge with respect to the image of the color signal in the device-independent color space, and the human visual frequency characteristic is added to the spatial frequency component. The value obtained by multiplying and integrating in the frequency domain and multiplying the obtained value by appropriate weighting can be used as the sharpness evaluation value Qs. If necessary, a value obtained by multiplying the obtained value by an appropriate weight can be used as the sharpness evaluation value Qs. The weight is set to an optimum value according to the image reading apparatus and image reading method used, the measured physical quantity or optical information, the color system, the visual characteristic correction function, or the correction method.

<Configuration of Tone / Color Reproducibility Evaluation Value Calculation Unit>
FIG. 5 is a block diagram illustrating a schematic configuration of the gradation / color reproducibility evaluation value calculation unit 260 of the image quality element evaluation value calculation unit 20. Here, a block diagram according to functional blocks of the processing procedure is shown. FIG. 6 is a diagram illustrating an example of the number density histogram of the target color area.

  The gradation / color reproducibility evaluation value calculation unit 260 calculates the gradation evaluation value Qt (Qt10, Qt20) and the color reproducibility evaluation value Qc (Qc10, Qc20) for each of the original image and the input image. Original image data D12 and D22 expressed by Lab-independent Lab values obtained by the color signal converter 16 are used. In calculating the gradation evaluation value Qt, the number density for an arbitrary color region in the device-independent color space is calculated, and the gradation evaluation value Qt is calculated from the standard deviation of the frequency distribution of the number density. It is configured. In addition, the color reproducibility evaluation value Qc is calculated by calculating the number density for an arbitrary color region in the device-independent color space and calculating the average value of the frequency distribution of the number density.

  Specifically, as shown in the drawing, the gradation / color reproducibility evaluation value calculation unit 260 calculates the number density in the target color region and the target color region setting unit 262 that sets the target color region. A reference grid setting unit 264 that sets a reference region and a number density distribution analysis unit 266 that calculates a frequency distribution (number density histogram) of the number density of Lab values in the target color region are provided.

  Further, the gradation / color reproducibility evaluation value calculation unit 260 obtains the gradation evaluation value Qt based on the standard deviation of the number density histogram analyzed by the number density distribution analysis unit 266. And an evaluation value calculation unit 270 having a color reproducibility evaluation value calculation unit 274 for obtaining a color reproducibility evaluation value Qc based on the average value of the number density histogram analyzed by the number density distribution analysis unit 266.

  For example, the target color area setting unit 262 sets an arbitrary target color area for which gradation or color reproducibility is to be evaluated in the Lab color space. For example, an area of a memory color such as a flesh color is set for a person image, and a blue sky or a grass color is set for a landscape photograph. In addition, the user can set the target color region while viewing the color solid that changes in brightness, saturation, and hue displayed on the display.

  The reference grid setting unit 264 sets a reference region when calculating the number density in the target color region set by the target color region setting unit 262. Here, the size of the reference lattice is, for example, a cube in which the Lab value is changed by 1 in the Lab space. However, the size of the reference grid is not limited to this.

  For example, as shown in FIG. 6, the number density distribution analysis unit 266 obtains a number density histogram of Lab values in the target color region. Here, the number density is the number of Lab values in the cube set by the reference grid setting unit 264.

  In the evaluation value calculation unit 270, the gradation evaluation value calculation unit 272 calculates the standard deviation of the number density histogram to obtain the gradation evaluation value Qt, and the color reproducibility evaluation value calculation unit 274 calculates the average value of the number density histogram. The color reproducibility evaluation value Qc.

  Here, the smaller the gradation evaluation value Qt, the better the gradation is because there is no gradation jump in the target color region. Further, the larger the color reproducibility evaluation value Qc, the more colors in the target color area and the better the color reproducibility.

  The above calculation process for calculating the evaluation value related to the gradation characteristic (gradation evaluation value Qt) and the evaluation value related to the color reproducibility (color reproducibility evaluation value Qc) is merely an example, and various changes are made. Any device can be used as long as it can cope with each image quality element. For example, the number density distribution analysis unit 266 obtains a brightness histogram, and an index (statistical) indicating the characteristics of the histogram (frequency distribution) such as an average value, median (median), or mode (mode) mode Index) may be calculated, and this index may be used as the gradation evaluation value Qt of the original image data D10. Similarly, the number density distribution analysis unit 266 obtains a saturation histogram or a hue histogram, calculates a statistical index indicating the characteristics of the histogram such as an average value, a median value, or a mode value, and uses the index as an original image. The color reproducibility evaluation value Qc of the data D10 may be used.

  Alternatively, the optical information difference between the adjacent gradation levels in the optical information representing the image is calculated, and the correction is added to the difference according to the human recognition limit for the optical information, and this corrected value is used for all gradation levels. The gradation evaluation value Qt may be obtained by integrating at. If necessary, the corrected values may be integrated with all gradation levels after being appropriately weighted. Corrections and weights according to the human visual perception limit are optimal depending on the image reading device and image reading method used, the measured physical quantity or optical information, the color system, the human perception limit deriving method and its value, etc. Set the value.

  As described above, the image quality element evaluation value calculation unit 20 inputs the image quality element evaluation values Qg, Qs, Qt, and Qc of graininess, sharpness, gradation, and color reproducibility with the original image data D10. It is directly calculated from the image data D20 (specifically, device-independent image data D12 and D22 color-converted by the color signal converter 16).

  Since the image quality element evaluation value calculation unit 20 calculates the image quality element evaluation values Qg, Qs, Qt, and Qc directly from the image data, the image quality element evaluation values Qg10, Qs10, Qt10, and Qc10 of the original image data D10. Includes image characteristics unique to the original image. For example, the granularity evaluation value decreases in an image including a lot of noise during image generation, and the sharpness evaluation value increases in an image including a large amount of image structure with high contrast. In addition, each of the image quality element evaluation values Qg20, Qs20, Qt20, and Qc20 of the input image data D20 indicates whether or not the image data is sent via the image output device 50 in acquiring the input image data D20 (for example, the original image data D10 is directly transmitted via the network). Regardless of (capture), the image characteristic unique to the original image is included, and when the input image data D20 is acquired by reading the image output by the image output apparatus 50, the output characteristic of the image output apparatus 50 is also included.

<Configuration of Output Image Quality Evaluation Value Calculation Unit>
FIG. 7 is a block diagram illustrating a schematic configuration of the output image quality evaluation value calculation unit 30. Here, a block diagram according to functional blocks of the processing procedure is shown.

  The output image quality evaluation value calculation unit 30 calculates the overall image quality evaluation value Q using a weighted linear regression equation of image quality element evaluation values. In this case, a linear equation using the product of the sharpness, the granularity and the gradation characteristic and the reciprocal of the sharpness as a variable is used, or the product of the sharpness, the granularity and the gradation characteristic constituting the linear equation, and the sharpness. The value of the reciprocal of the degree is corrected to a power by a predetermined constant. The coefficient is adjusted (to a different value) according to the image type.

  In any case, the granularity evaluation value Qg (Qg10, Qg20), the sharpness evaluation value Qs (Qs10, Qs20), and the gradation property for each of the original image and the input image obtained by the image quality element evaluation value calculation unit 20. Using the evaluation value Qt (Qt10, Qt20) and the color reproducibility evaluation value Qc (Qc10, Qc20), an overall image quality evaluation value Q representing the overall image quality felt by humans with a good correlation with human senses is obtained.

  Specifically, as shown in the figure, the output image quality evaluation value calculation unit 30 outputs an output characteristic analysis unit 300 that analyzes the image quality characteristic (output characteristic) of the image output device 50 and an analysis result by the output characteristic analysis unit 300. And an overall image quality evaluation value calculation unit 320 that calculates an overall image quality evaluation value Q of the image output apparatus 50 that can correspond to the subjective overall image quality.

  The output characteristic analysis unit 300 outputs output characteristics (granularity) of the image output device 50 based on the granularity evaluation values Qg10 and Qg20 for the original image and the input image obtained by the granularity evaluation value calculation unit 220, respectively. The image output device 50 based on the sharpness evaluation values Qs10 and Qs20 for the original image and the input image obtained by the graininess analysis unit 302 for analyzing the image quality evaluation value) and the sharpness evaluation value calculation unit 240, respectively. A sharpness analysis unit 304 that analyzes output characteristics (sharpness evaluation value) relating to sharpness.

  In addition, the output characteristic analysis unit 300 has a gradation evaluation value Qt10, for each of the original image and the input image obtained by the gradation evaluation value calculation unit 272 of the gradation / color reproducibility evaluation value calculation unit 260. Based on Qt20, the color reproducibility of the gradation characteristic analysis unit 306 that analyzes the output characteristic (gradation characteristic evaluation value) related to the gradation characteristic of the image output device 50 and the gradation characteristic / color reproducibility evaluation value calculation part 260 Based on the color reproducibility evaluation values Qc10 and Qc20 for the original image and the input image obtained by the evaluation value calculation unit 274, the output characteristics (color reproducibility evaluation value) regarding the color reproducibility of the image output device 50 are analyzed. A color reproducibility analysis unit 308.

  The image quality element evaluation value calculation unit 20 directly calculates the evaluation values Q10 and Q20 from the image data D10 (preferably original image data D12) and D20 (preferably input image data D22), and the output image image quality evaluation value calculation unit 30 By analyzing the output characteristics of the image output device 50 by comparing the evaluation values Q10 and Q20, the evaluation due to the misalignment between the two images that may occur by taking the difference between the pixels of the input image and the original image Eliminates the problem of reduced accuracy.

  Specifically, each of the analysis units 302, 304, 306, and 308 includes image quality element evaluation values Qg10, Qg20, Qs10, Qs20, Qt10, Qt20, Qc10, and Qc20 corresponding to the original image data D10 and the input image data D20. By calculating the difference, the granularity evaluation value QgΔ, the sharpness evaluation value QsΔ, the gradation evaluation value QtΔ, and the color reproducibility evaluation value QcΔ representing the image quality of the output image are analyzed. The difference between the image quality element evaluation values can be used as an evaluation value representing the quality (output characteristics) of the output image of the image output device 50.

  The overall image quality evaluation value calculation unit 320 converts the granularity evaluation value QgΔ, sharpness evaluation value QsΔ, gradation evaluation value QtΔ, and color reproducibility evaluation value QcΔ of the image output device 50 obtained by the output characteristic analysis unit 300. Based on this, an overall image quality evaluation value Q corresponding to the subjective overall image quality is calculated.

  For example, a weighting coefficient of each image quality element evaluation value for subjective overall image quality is obtained in advance from a subjective evaluation experiment, and the overall image quality evaluation value Q is calculated according to a weighted linear sum of image quality element evaluation values shown in the following equation (11). calculate. In the equation, kg is a weighting factor for the granularity evaluation value, ks is a weighting factor for the sharpness evaluation value, kt is a weighting factor for the gradation characteristic evaluation value, and kc is a weighting factor for the color reproducibility evaluation value.

ここで、本実施形態では、出力特性解析部３００において、原画像および入力画像のそれぞれから直接に算出した画質要素評価値Ｑｇ10，Ｑｇ20、Ｑｓ10，Ｑｓ20，Ｑｔ10，Ｑｔ20，Ｑｃ10，Ｑｃ20の差分を取り総合画質評価値Ｑを算出しているので、従来人間の目視によって主観的に行なわれていた画像の総合的な画像品質評価を定量的かつ安定的に行なうことができ、また、心理物理量を組み合わせた量を変数としているので、人間の感覚と相関がよく高精度な総合画質評価値Ｑを算出することができる。

Here, in this embodiment, the output characteristic analysis unit 300 takes the difference between the image quality element evaluation values Qg10, Qg20, Qs10, Qs20, Qt10, Qt20, Qc10, Qc20 directly calculated from the original image and the input image. Since the total image quality evaluation value Q is calculated, it is possible to quantitatively and stably perform comprehensive image quality evaluation of images that have been subjectively performed by human eyes, and to combine psychophysical quantities. Therefore, it is possible to calculate the overall image quality evaluation value Q having a good correlation with human sense and high accuracy.

  Further, the evaluation value calculation unit 270 does not need to take a difference for each pixel between the two images, can ignore the influence of the positional deviation between the input image and the original image, and can calculate the image evaluation value accurately and easily.

  In addition, since the evaluation value calculation unit 270 calculates the difference between the evaluation values of the original image and the input image, only the output characteristics of the image output device 50 can be obtained even when the image quality of the original image is poor or the pattern of the original image is different. Can be calculated.

  The output image quality evaluation value calculation unit 30 compares the image quality element evaluation values Qg10, Qg20, Qs10, Qs20, Qt10, Qt20, Qc10, and Qc20 of the original image data D10 and the input image data D20 to output an image. It is only necessary to obtain an index value indicating the output characteristics of the device 50, that is, a graininess evaluation value, a sharpness evaluation value, a gradation evaluation value, and a color reproducibility evaluation value. For example, the ratio may be used as an index value indicating output characteristics.

<Correlation with subjective evaluation score>
FIG. 8 is a diagram illustrating an example of the correlation between the overall image quality evaluation value Q and the subjective evaluation score obtained from the subjective evaluation experiment when a person image is used as the original image data D10. In addition, the weighting coefficient of each evaluation value is set to kg = 0.288, ks = 0.326, kt = 0.088, kc = 0.298, and kn = 0, respectively.

  As is clear from the equations (3-1), (3-2), (3-3) and the equations (5-1), (5-2), (5-3), the overall image quality according to the present embodiment. The correlation between the evaluation value Q and the subjective evaluation score is very high, and the contribution rate is 0.906.

  In this way, by calculating the overall image quality evaluation value Q by multiplying the image quality element evaluation value by the weighting coefficient obtained from the subjective evaluation experiment, it is possible to calculate an evaluation value corresponding to the human overall image quality.

  As described above, according to the image quality evaluation method by the image evaluation apparatus 10 of the present embodiment, the quality of the image output from the image output apparatus 50 can be accurately and easily evaluated.

  Also, according to the image quality evaluation method of the present embodiment, when evaluating the image quality of the image output from the image output device 50, the difference between the image quality element evaluation values calculated directly from the image is taken to obtain the overall image quality evaluation value Q. Since the calculation is performed, it is not necessary to take a difference for each pixel, the influence of the positional deviation between the input image and the original image can be ignored, and the evaluation value can be calculated accurately and easily.

  In addition, when calculating the evaluation value, since the color signal of the device-independent color space is used and the evaluation result obtained from the subjective evaluation experiment is taken into consideration in advance, The obtained evaluation value can be calculated.

  In addition to pattern images such as patch images and gradation images, pattern images such as photographic images can be used, so the color conversion processing depending on the pattern image is performed by the automatic image quality adjustment function of the printer driver. Thus, the image quality evaluation value can be accurately calculated for the output image.

  In addition, since the evaluation value calculation unit 270 calculates the difference between the image quality element evaluation values of the input image and the original image, an image quality evaluation value that does not depend on the image quality or the design of the original image can be calculated. Even when the pattern of the image is different, the image quality evaluation value representing only the output characteristic of the image output device 50 can be calculated.

<Configuration using an electronic computer>
The series of image evaluation processing functions described above can be executed by hardware, but can also be executed by software. In other words, the control unit and each functional unit related to the execution of the image evaluation processing are not limited to being configured by hardware, but are realized by software using a computer (computer) based on a program code that realizes the function. It is also possible to do. For example, similarly to the configuration of a general PC (personal computer) main body, a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read-Only Memory), and the like can be provided.

  The execution of the image evaluation process described above can be realized by the CPU executing an application program incorporated in the image evaluation apparatus having the computer configuration. Therefore, an image quality evaluation program suitable for realizing the image evaluation method and apparatus according to the present invention by software using an electronic computer (computer) or a computer-readable storage medium storing the image quality evaluation program is extracted as an invention. You can also

  Of course, the configuration is not limited to such a computer, but may be configured by a combination of dedicated hardware that performs each function. By adopting a mechanism for executing processing by software, it is possible to enjoy the advantage that the processing procedure, processing conditions, and the like can be easily changed without changing hardware. On the contrary, if it is configured by hardware, high-speed processing can be realized.

  The recording medium causes a change state of energy such as magnetism, light, electricity, etc. to the reading device provided in the hardware resource of the computer according to the description content of the program, and in the form of a signal corresponding thereto. The program description can be transmitted to the reader.

  For example, a magnetic disc (including a flexible disc), an optical disc (CD-ROM (Compact Disc-Read Only Memory)), a DVD (which is distributed to provide a program to a user separately from a computer, (Including Digital Versatile Disc), magneto-optical disc (including MD (Mini Disc)), or package media (portable storage media) made of semiconductor memory, etc. It may be configured by a ROM, a hard disk, or the like in which a program is recorded that is provided to the user in a state. Or the program which comprises software may be provided via communication networks, such as a wire communication or radio | wireless.

  When a series of evaluation value calculation processes are executed by software, a program (such as an embedded microcomputer) in which a program constituting the software is incorporated in dedicated hardware, or a CPU, logic circuit, storage device, etc. System on a chip (SOC) that implements a desired system by mounting functions on a single chip, or general-purpose that can execute various functions by installing various programs It is installed from a recording medium in a personal computer or the like.

  For example, a storage medium in which a program code of software for realizing an image evaluation value calculation processing function is supplied to a system or apparatus, and the computer (or CPU or MPU) of the system or apparatus stores the program code in the storage medium The same effect as that achieved by hardware can also be achieved by reading and executing. In this case, the program code itself read from the storage medium realizes an image evaluation value calculation processing function.

  Further, by executing the program code read by the computer, not only the function of performing the image evaluation value calculation process is realized, but also an OS (operating system; operating on the computer) based on the instruction of the program code. Basic software) or the like may perform a part or all of the actual processing, and the function of performing the image evaluation value calculation processing may be realized by the processing.

  Further, after the program code read from the storage medium is written in a memory provided in a function expansion card inserted into the computer or a function expansion unit connected to the computer, the function expansion is performed based on the instruction of the program code. There may be a case where the CPU or the like provided in the card or the function expansion unit performs part or all of the actual processing, and the function of performing the image evaluation value calculation processing is realized by the processing.

  In such an electronic computer, for example, software similar to that in a conventional image forming apparatus (multifunction machine) such as a copying application, a printer application, a facsimile (FAX) application, or a processing program for other applications is incorporated. . A control program for transmitting and receiving data to and from the outside via the network 9 is also incorporated.

  At this time, the program is provided as a file describing a program code for realizing the function of performing the image evaluation value calculation process. In this case, the program is not limited to being provided as a batch program file, and is a system configured by a computer. Depending on the hardware configuration, it may be provided as an individual program module. For example, it may be provided as add-in software incorporated in existing copying apparatus control software or printer control software (printer driver).

  FIG. 9 shows an example of a hardware configuration in which the image evaluation apparatus 10 is configured by software using a CPU and a memory, that is, the image evaluation apparatus 10 is realized by software using an electronic computer such as a personal computer. FIG.

  The computer system 900 constituting the image evaluation apparatus 10 includes data from a controller unit 901 and a predetermined storage medium such as a hard disk device, a flexible disk (FD) drive, a CD-ROM (Compact Disk ROM) drive, or a semiconductor memory controller. And a recording / reading control unit 902 for reading out and recording data. In particular, in the present embodiment, the hard disk device or other storage medium functions as the original image storage unit 12 or the input image storage unit 14 to store the original image data D10 and the like.

  The controller unit 901 includes a CPU (Central Processing Unit) 912, a ROM (Read Only Memory) 913 which is a read-only storage unit, and a RAM (Random Access) which can be written and read at any time and is an example of a volatile storage unit. Memory) 915 and RAM (described as NVRAM) 916 which is an example of a nonvolatile storage unit.

  In the above description, the “volatile storage unit” means a storage unit in a form in which the stored contents are lost when the power of the image output terminal 4 is turned off. On the other hand, the “nonvolatile storage unit” means a storage unit in a form that keeps stored contents even when the main power supply of the image output terminal 4 is turned off. Any memory device can be used as long as it can retain the stored contents. The semiconductor memory device itself is not limited to a nonvolatile memory device, and a backup power supply is provided to make a volatile memory device “nonvolatile”. You may comprise as follows. Further, the present invention is not limited to a semiconductor memory element, and may be configured using a medium such as a magnetic disk or an optical disk.

  The computer system 900 also includes an instruction input unit 903 having a keyboard, a mouse, and the like as a function unit that forms a user interface, and a display output unit 904 that presents a user with predetermined information such as a guidance screen and a processing result during operation. An image reading unit (scanner unit) 905 that reads an image to be processed, an image forming unit 906 that outputs a processed image in the image output terminal 4 to a predetermined output medium (for example, printing paper), and each functional unit And an interface unit 909 that performs an interface function therebetween.

  Examples of the interface unit 909 include a system bus 991 that is a transfer path of processing data (including image data) and control data, a scanner IF unit 995 that functions as an interface with the image reading unit 905, an image forming unit 906, and the like. It has a printer IF unit 996 that functions as an interface with other printers, and a communication IF unit 999 that mediates transfer of communication data with the network 9 such as the Internet. The scanner IF unit 995 and the communication IF unit 999 correspond to the input interface unit 11in, and the printer IF unit 996 corresponds to the output interface unit 11out. Note that the recording / reading control unit 902 also has a function of the input interface unit 11in that captures image data from a predetermined storage medium.

  The display output unit 904 includes a display device for presenting main information such as the read whole image and guidance information. The display device is preferably arranged in the vicinity of the instruction input unit 903 so as to efficiently perform a predetermined input operation while confirming the displayed information.

  The display output unit 904 includes, for example, a display control unit 942 and a display unit 944 made up of a CRT (Cathode Ray Tube) or LCD (Liquid Crystal Display). For example, the display control unit 942 displays guidance information, the entire image captured by the image reading unit 905, and the like on the display unit 944. Note that by using the display unit 944 having the touch panel 932 on the display surface, the instruction input unit 903 for inputting predetermined information with a fingertip or a pen can be configured.

  The computer system 900 is connected as an image reading unit corresponding to the image input device 70 of the above embodiment. The image input device 70 has the function of an image input terminal. For example, by using a full width array of CCD solid-state imaging devices, the image input device 70 irradiates light on the document sent to the reading position, thereby displaying an image on the document. Read and convert analog video signals of red R, green G, and blue B representing the read image into digital signals. In particular, in the present embodiment, the read image is stored as input image data D20 in a hard disk device or other storage medium (or RAM 915).

  The computer system 900 is connected to an image forming unit corresponding to the image output device 50 of the above embodiment. This image output device 50 is an electrophotographic type, thermal type, thermal transfer type, ink jet type or similar conventional image obtained by, for example, representing an image represented by an image signal obtained by an image reading unit (image input device 70). Using the forming process, a visible image is formed (printed) on plain paper or thermal paper.

  Therefore, the image forming unit (image output device 50) includes, for example, an image processing unit 52 that generates print output data such as binary signals of yellow Y, magenta M, cyan C, and black K, and the image evaluation device 10. Is equipped with a raster output scan-based print engine 54 for operating as a digital printing system.

  In such a configuration, the CPU 912 controls the entire system via the system bus 991. The ROM 913 stores a control program for the CPU 912 and the like. The RAM 915 is configured by SRAM (Static Random Access Memory) or the like, and stores program control variables, data for various processes, and the like. In addition, the RAM 915 stores an electronic document (not only character data but also image data) acquired by a predetermined application program, and image data acquired by an image reading unit (image input device 70) provided in the own device. In addition, an area for temporarily storing electronic data obtained from the outside is included.

  For example, a program that causes a computer to execute an image evaluation value calculation process is distributed through a recording medium such as a CD-ROM. Alternatively, this program may be stored in the FD instead of the CD-ROM. In addition, an MO drive may be provided to store the program in the MO, or the program may be stored in another recording medium such as a nonvolatile semiconductor memory card such as a flash memory. Furthermore, the program may be downloaded and acquired from another server or the like via the network 9 such as the Internet, or may be updated.

  As a recording medium for providing the program, in addition to FD and CD-ROM, an optical recording medium such as DVD, a magnetic recording medium such as MD, a magneto-optical recording medium such as PD, a tape medium, and a magnetic medium. A semiconductor memory such as a recording medium, an IC card, or a miniature card can be used. Part or all of the image evaluation value calculation processing function in the image evaluation apparatus 10 can be stored in an FD or CD-ROM as an example of a recording medium.

  Also, the hard disk device stores data for various processes by the control program, temporarily stores a large amount of image data acquired by the image reading unit (image input device 70), print data acquired from the outside, and the like. Or the area to be. The hard disk device, FD drive, or CD-ROM drive is used, for example, for registering program data for causing the CPU 912 to execute processing such as content acquisition, address acquisition, or address setting by software. .

  It should be noted that a processing circuit 908 may be provided in which not all processing of each functional part related to the image evaluation value calculation processing of the image evaluation apparatus 10 is performed by software, but a part of these functional parts is performed by dedicated hardware. Good. In this case, as illustrated, the color signal conversion unit 984 corresponding to the color signal conversion unit 16, the image quality element evaluation value calculation unit 986 corresponding to the image quality element evaluation value calculation unit 20, and the output image image quality evaluation value calculation unit 30 correspond to. A comprehensive image quality evaluation value calculation unit 988 or the like may be provided individually.

  Although the mechanism performed by software can flexibly cope with parallel processing and continuous processing, the processing time becomes longer as the processing becomes complicated, so that a reduction in processing speed becomes a problem. On the other hand, it is possible to construct an accelerator system with a higher speed by using a hardware processing circuit. Even if the processing is complicated, the accelerator system can prevent a reduction in processing speed and can calculate a high throughput.

  As mentioned above, although this invention was demonstrated using embodiment, the technical scope of this invention is not limited to the range as described in the said embodiment. Various changes or improvements can be added to the above-described embodiment without departing from the gist of the invention, and embodiments to which such changes or improvements are added are also included in the technical scope of the present invention.

  Further, the above embodiments do not limit the invention according to the claims (claims), and all combinations of features described in the embodiments are not necessarily essential to the solution means of the invention. Absent. The embodiments described above include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. Even if some constituent requirements are deleted from all the constituent requirements shown in the embodiment, as long as an effect is obtained, a configuration from which these some constituent requirements are deleted can be extracted as an invention.

  In the above-described embodiment, the evaluation method of the image output from the image output device 50 such as a color printer has been described as an example. However, the output is performed not only from the color printer but also from other image output devices such as a display, electronic paper, and a projector. The mechanism described in the above embodiment can be similarly applied to the evaluation of the image quality of a color image.

It is a figure explaining the example of 1 composition of the image processing system provided with the image evaluation device concerning the present invention. It is a block diagram explaining the schematic structure of a color signal conversion part. It is a block diagram explaining the schematic structure of a granularity evaluation value calculation part. It is a block diagram explaining the schematic structure of a sharpness evaluation value calculation part. It is a block diagram explaining the schematic structure of a gradation property / color reproducibility evaluation value calculation part. It is a figure which shows an example of the number density histogram of an attention color area. It is a block diagram explaining the schematic structure of a comprehensive image quality evaluation value calculation part. It is a figure which shows an example of correlation with the comprehensive image quality evaluation value at the time of using a person image, and the subjective evaluation score obtained from the subjective evaluation experiment. It is the figure which showed an example of the hardware constitutions in the case of implement | achieving an image evaluation apparatus like software using electronic computers, such as a personal computer.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Image evaluation system, 10 ... Image evaluation apparatus, 11out ... Output interface part, 11in ... Input interface part, 12 ... Original image memory | storage part, 14 ... Input image memory | storage part, 16 ... Color signal conversion part, 20 ... Image quality element evaluation Value calculation unit, 30 ... output image quality evaluation value calculation unit, 50 ... image output device, 70 ... image input device, 162 ... original image data color conversion unit, 164 ... input image data color conversion unit, 220 ... granularity evaluation value Calculation unit 222 ... Color difference calculation unit, 224 ... Two-dimensional orthogonal transformation unit, 226 ... Two-dimensional power spectrum calculation unit, 228 ... Two-dimensional power spectrum one-dimensionalization calculation unit, 230 ... Luminous spatial frequency characteristic correction unit, 232 ... granularity evaluation value calculation unit, 240 ... sharpness evaluation value calculation unit, 242 ... edge amount calculation unit, 244 ... edge strength calculation unit, 246 ... sharpness evaluation value calculation unit, 260 ... gradation Color reproducibility evaluation value calculation unit, 262... Attention color region setting unit, 264... Reference grid setting unit, 266 .. number density distribution analysis unit, 270... Evaluation value calculation unit, 272. ... color reproducibility evaluation value calculation unit, 300 ... output characteristic analysis unit, 302 ... graininess analysis unit, 304 ... sharpness analysis unit, 306 ... gradation property analysis unit, 308 ... color reproducibility analysis unit, 320 ... total image quality Evaluation value calculator

Claims (16)

An original image quality element evaluation value calculation unit for calculating an original image quality element evaluation value representing the image quality of the output image based on original image data representing the output image;
An input interface unit that captures input image data corresponding to the original image data;
An input image quality element evaluation value calculation unit that calculates an input image quality element evaluation value representing an image quality of an input image corresponding to the output image based on the input image data captured by the input interface unit;
By comparing the original image quality element evaluation value calculated by the original image quality element evaluation value calculation unit with the input image quality element evaluation value calculated by the input image quality element evaluation value calculation unit, the output image An image evaluation apparatus comprising: an output image quality evaluation value calculation unit that calculates an output image quality evaluation value representing image quality. The input interface unit captures the input image data from an input image data generation unit that optically reads an output image formed based on the original image data and generates the input image data. 2. The image evaluation apparatus according to 1. An input characteristic correction unit that corrects input characteristics of the input image data generation means for generating the input image data;
Each of the said image quality element evaluation value calculation part calculates each said evaluation value showing image quality based on the image data correct | amended by the said input characteristic correction | amendment part. Image evaluation device. A color signal conversion unit that converts image data representing an image into a color signal in a device-independent color space;
Each of the said image quality element evaluation value calculation part calculates each said evaluation value showing image quality based on the color signal color-converted by the said color signal conversion part. The image evaluation apparatus according to any one of the above. The original image quality element evaluation value calculation unit and the input image image quality element evaluation value calculation unit relate to at least one of image graininess, sharpness, gradation characteristics, and color reproducibility based on the image data. Find each evaluation value,
The output image image quality evaluation value calculation unit calculates the output image image quality evaluation value by comparing the obtained corresponding evaluation values with each other. The image evaluation apparatus according to 1. The original image image quality element evaluation value calculation unit and the input image image quality element evaluation value calculation unit are configured to output a plurality of output characteristics among image graininess, sharpness, gradation characteristics, and color reproducibility based on the image data. Each evaluation value for
The output image image quality evaluation value calculation unit obtains an evaluation value relating to each image quality characteristic by comparing the obtained corresponding evaluation values, and further, based on the evaluation value relating to the individual image quality characteristic, The image evaluation apparatus according to claim 5, wherein an overall image quality evaluation value representing the overall image quality is calculated, and the image quality is evaluated based on the overall evaluation value. The original image quality element evaluation value calculation unit and the input image image quality element evaluation value calculation unit calculate a spatial frequency component of the processing target image, and according to the spatial frequency characteristic of the visual system with respect to the calculated spatial frequency component The image evaluation apparatus according to claim 5, wherein each evaluation value relating to the graininess of the processing target image is obtained by performing correction and integrating the corrected spatial frequency component. The original image quality factor evaluation value calculation unit and the input image quality factor evaluation value calculation unit perform two-dimensional orthogonal transformation on the processing target image to generate information indicating a two-dimensional spatial frequency distribution. The image evaluation apparatus according to claim 7, wherein after the dimensionalization, correction according to a spatial frequency characteristic of the visual system is performed. The original image quality element evaluation value calculation unit and the input image image quality element evaluation value calculation unit calculate a spatial frequency component of the processing target image, and each evaluation regarding the sharpness of the processing target image based on the calculated spatial frequency component The image evaluation apparatus according to claim 5, wherein a value is obtained. The original image quality element evaluation value calculation unit and the input image image quality element evaluation value calculation unit apply a differential filter to the processing target image, so that a horizontal frequency and a vertical direction as a spatial frequency component of the processing target image are obtained. The image evaluation apparatus according to claim 9, wherein each edge amount is calculated, and each evaluation value relating to the sharpness of the processing target image is obtained from an average of a square root of a square sum of each edge amount. The original image quality element evaluation value calculation unit and the input image image quality element evaluation value calculation unit calculate the number density of pixel information in an arbitrary color region of the processing target image, and based on the standard deviation of the frequency distribution of the number density The image evaluation apparatus according to claim 5, wherein each evaluation value related to a gradation characteristic of the processing target image is obtained. The original image quality element evaluation value calculation unit and the input image image quality element evaluation value calculation unit calculate the number density of pixel information in an arbitrary color region of the processing target image, and based on an average value of the frequency distribution of the number density The image evaluation apparatus according to claim 5, wherein each evaluation value relating to color reproducibility of the processing target image is obtained. The output image image quality evaluation value calculation unit includes the graininess, sharpness, gradation characteristics, and color reproducibility obtained by the original image image quality element evaluation value calculation unit and the input image image quality element evaluation value calculation unit. The image evaluation apparatus according to claim 6, wherein the overall image quality evaluation value is calculated by a predetermined linear equation using each evaluation value relating to the plurality of output characteristics. The output image image quality evaluation value calculation unit calculates the overall image quality evaluation value using a weighted linear regression equation obtained by multiplying each evaluation value by a weighting coefficient corresponding to an image type as the linear equation. The image evaluation apparatus according to claim 13. Calculating an original image quality element evaluation value representing the image quality of the output image based on the original image data representing the output image;
Capturing input image data corresponding to the original image data;
Calculate an input image quality element evaluation value representing the image quality of the input image corresponding to the output image based on the captured input image data,
An output image quality evaluation value representing an image quality of an output image is calculated by comparing the original image quality element evaluation value calculated in this way and the input image quality element evaluation value. Method. A program for performing image evaluation processing using a computer to evaluate the image quality of an image to be evaluated using a computer,
The computer,
An original image quality element evaluation value calculation unit for calculating an original image quality element evaluation value representing the image quality of the output image based on original image data representing the output image;
An input image quality element evaluation value calculation unit that takes in input image data corresponding to the original image data and calculates an input image quality element evaluation value representing the image quality of the input image corresponding to the output image based on the input image data When,
By comparing the original image quality element evaluation value calculated by the original image quality element evaluation value calculation unit with the input image quality element evaluation value calculated by the input image quality element evaluation value calculation unit, the output image A program that functions as an output image quality evaluation value calculation unit that calculates an output image quality evaluation value representing image quality.

Images