JP2016066990A - Evaluation method of tone correction, and arithmetic device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an evaluation method of tone correction which allows for appropriate evaluation of tone correction performed for an image, and to provide an arithmetic device.SOLUTION: In a method for evaluating the effect of tone correction for an image, pixels included in an image to be corrected is classified to a dark part, an intermediate part, and a bright part, of a plurality of luminance value levels, depending on the luminance thereof, the pixels included in a corrected image are classified to have the same luminance value level as that of a pixel at the same position of the image to be corrected, statistic value of the image to be corrected and statistic value of the corrected image are determined by using the luminance value of each pixel of a pixel group included in each luminance value level, an evaluation index is determined using statistic value of the image to be corrected and statistic value of the corrected image, and then tone correction is evaluated using the evaluation index.SELECTED DRAWING: Figure 4

Description

  The present invention relates to a technique for evaluating the effect of gradation correction performed on an image.

  Conventionally, an image quality is evaluated for an image. In particular, there has been proposed a technique for evaluating the clearness with a good correlation with a human sense when input is performed through a device (see Patent Document 1). In addition to simply evaluating the sharpness of an image, various gradation corrections such as contrast correction (brightness correction) are also performed on the image in order to give sharpness to an image with low sharpness. Yes.

  In particular, visibility can be improved by applying contrast correction using a tone curve or the like when a captured image becomes dark according to the settings at the time of shooting. In the tone curve, the luminance value before and after correction is defined by a line graph, and the luminance value after correction can be adjusted by deforming the tone curve. The luminance value includes not only a value that expresses the luminance purely but also various values that express the gradation by the original value of the pixel. In general, correction according to a uniform tone curve is applied to the entire image, but bright and dark portions exist in the image by appropriately updating the slope of the tone curve according to the luminance value of the original image. Even in this case, an image with good visibility can be obtained. Various methods have been proposed for such contrast correction.

JP 2000-188647 A

  However, as shown in the above-mentioned Patent Document 1, evaluation of the sharpness of the image itself has been performed, but evaluation of the effect of gradation correction on the image has not been performed.

  Accordingly, it is an object of the present invention to provide a gradation correction evaluation method capable of appropriately evaluating the effect of gradation correction performed on an image, and an arithmetic device for evaluation of gradation correction. And

In order to solve the above problems, in the first aspect of the present invention,
Obtain the statistical value of the pixel value of the pre-correction image that is the image before the gradation correction and the statistical value of the pixel value of the corrected image that is the image after the gradation correction,
Using the statistical value of the pixel value of the image before correction and the statistical value of the pixel value of the image after correction, an evaluation index is obtained,
A gradation correction evaluation method is provided, wherein gradation evaluation is evaluated using the evaluation index.

  According to the first aspect of the present invention, the statistical value of the pixel value of the image before correction and the statistical value of the pixel value of the image after correction are obtained, and the statistical value of the pixel value of the image before correction and the pixel value of the image after correction are calculated. Since the evaluation index is obtained using the statistical value and the gradation correction is evaluated using the obtained evaluation index, the effect of the gradation correction performed on the image can be appropriately evaluated. Become.

  In the second aspect of the present invention, in the gradation correction evaluation method, a set of pixels satisfying predetermined conditions is selected as a selected pixel group from the uncorrected image and the corrected image, and the selected pixel is selected. A pixel value of a pixel belonging to the group is acquired, and the statistical value is obtained based on the pixel value. According to the second aspect of the present invention, a set of pixels satisfying a predetermined condition is selected as a selected pixel group from the pre-correction image and the post-correction image, and before gradation correction is performed on the pixels belonging to the selected pixel group. And the corrected pixel value is obtained, and the statistical value is obtained based on the pixel value, so that the effect of gradation correction not only on the entire image but also on the pixel group having a specific feature should be evaluated appropriately. Is possible.

  In the third aspect of the present invention, in the gradation correction evaluation method, a plurality of the selected pixel groups exist for each of the pre-correction image and the post-correction image, and pixel values of pixels belonging to each selected pixel group are determined. Obtaining and obtaining the statistical value based on the pixel value. According to the third aspect of the present invention, the pixel values before and after the gradation correction are obtained for the pixels belonging to the plurality of selected pixel groups, and the statistical values are obtained. It is possible to appropriately evaluate the effect of gradation correction for each pixel group.

  In the fourth aspect of the present invention, in the gradation correction evaluation method, the predetermined condition is a range of pixel values. According to the fourth aspect of the present invention, since the selected pixel group is specified by the range of pixel values, the effect of gradation correction is individually evaluated for pixel value levels such as a bright portion, an intermediate portion, and a dark portion. It becomes possible to do.

  In the fifth aspect of the present invention, in the gradation correction evaluation method, the predetermined condition is that the predetermined condition is included in a partial image that is a part of the pre-correction image and the post-correction image. And According to the fifth aspect of the present invention, since the selected pixel group is specified by the partial image that is a part of the pre-correction image and the post-correction image, the effect of gradation correction is individually determined according to the position on the image. Can be evaluated.

  In the sixth aspect of the present invention, in the gradation correction evaluation method, the partial image is specified by an area obtained by dividing the pre-correction image and the post-correction image in a grid pattern. According to the sixth aspect of the present invention, since the partial image is an image obtained by dividing the pre-correction image and the post-correction image in a grid pattern, the effect of gradation correction can be individually evaluated for the entire area of the image. It becomes possible.

In the seventh aspect of the present invention,
The pixels included in the pre-correction image that is the image before gradation correction are classified into a plurality of pixel value levels according to the pixel values, and the pixels included in the corrected image that is the image after gradation correction are Classify the same pixel value level as the pixel at the same position in the pre-correction image,
For the pre-correction image and the post-correction image, obtain a statistical value of the pixel value of each pixel of the pixel group included in each pixel value level,
Using the statistical value of the pixel value of the image before correction and the statistical value of the pixel value of the image after correction, an evaluation index is obtained,
A gradation correction evaluation method is provided, wherein gradation evaluation is evaluated using the evaluation index.

  According to the seventh aspect of the present invention, the pixels included in the pre-correction image are classified into a plurality of pixel value levels according to the pixel values, and the pixels included in the post-correction image that is the image after gradation correction are determined. The pixel value level is classified into the same pixel value level as the pixel at the same position in the pre-correction image, and the statistical value of the pixel value of each pixel of the pixel group included in each pixel value level is obtained for the pre-correction image and the post-correction image. Since the evaluation index is obtained using the statistical value of the pixel value of the image and the statistical value of the pixel value of the corrected image, and the gradation correction is evaluated using the obtained evaluation index, the bright part of the image, It is possible to individually evaluate the effect of gradation correction for a plurality of pixel value levels such as a luminance value level such as an intermediate portion and a dark portion.

  In the eighth aspect of the present invention, in the gradation correction evaluation method, the range of the pixel value is determined by a threshold value set in a portion that becomes a valley of the histogram. According to the eighth aspect of the present invention, the range of pixel values specifying the selected pixel group is determined by the threshold value set in the portion that becomes the valley of the histogram. It is possible to optimize the division.

In the ninth aspect of the present invention,
The pixel value of each pixel of the pixel group included in the partial image obtained by dividing the pre-correction image, which is an image before gradation correction, and the post-correction image, which is an image after gradation correction, into a grid pattern, before the correction Obtain the statistical value of the pixel value of the image and the statistical value of the pixel value of the corrected image,
Using the statistical value of the pixel value of the image before correction and the statistical value of the pixel value of the image after correction, an evaluation index is obtained,
A gradation correction evaluation method is provided, wherein gradation evaluation is evaluated using the evaluation index.

  According to the ninth aspect of the present invention, using the pixel value of each pixel of the pixel group included in the partial image obtained by dividing the pre-correction image and the post-correction image in a grid pattern, , Calculate the statistical value of the pixel value of the corrected image, determine the evaluation index using the statistical value of the pixel value of the image before correction and the statistical value of the pixel value of the image after correction, and then use the calculated evaluation index to Since the correction is evaluated, the effect of the gradation correction can be individually evaluated according to the position on the image.

  The tenth aspect of the present invention is characterized in that, in the gradation correction evaluation method, the partial image is classified by a pixel value range in a pixel group unit arranged in a rectangular shape. To do. According to the tenth aspect of the present invention, the partial images that are a part of the pre-correction image and the post-correction image are classified by the range of pixel values in units of pixels arranged in a rectangular shape. It is possible to individually evaluate the effect of gradation correction in accordance with the position on the image and the pixel value levels such as the bright part, the intermediate part, and the dark part.

  In the eleventh aspect of the present invention, in the gradation correction evaluation method, the partial image is determined based on a designated position on the image. According to the eleventh aspect of the present invention, the partial image that is a part of the pre-correction image and the post-correction image is determined based on the designated position on the image. It is possible to acquire an optimal partial image that satisfies the conditions.

  In the twelfth aspect of the present invention, in the gradation correction evaluation method, the evaluation index is a ratio of a statistical value of the pixel value of the image before correction and a statistical value of the pixel value of the image after correction. Features. According to the twelfth aspect of the present invention, since the evaluation index is the ratio between the statistical value of the pixel value of the pre-correction image and the statistical value of the pixel value of the post-correction image, the effect of gradation correction is more appropriately evaluated. It becomes possible.

  In the thirteenth aspect of the present invention, in the gradation correction evaluation method, the statistical value is a value based on a standard deviation. According to the thirteenth aspect of the present invention, since the statistical value is a value based on the standard deviation, it is possible to more appropriately evaluate the effect of the gradation correction in consideration of the average and variance of the pixel values.

  In the fourteenth aspect of the present invention, the statistical value is a value based on a difference between a maximum value and a minimum value of pixel values. According to the fourteenth aspect of the present invention, since the statistical value is a value based on the difference between the maximum value and the minimum value of the pixel value, the gradation correction effect is more appropriately evaluated in consideration of the maximum range of the pixel value. It becomes possible to do.

  In the fifteenth aspect of the present invention, the statistical value is a value based on a difference between the maximum value and the minimum value of the specified pixel value by specifying a pixel whose pixel value is included in the central predetermined ratio. It is characterized by. According to the fifteenth aspect of the present invention, the statistical value is a value based on the difference between the maximum value and the minimum value of the pixel values that are included in the central predetermined ratio, so that the pixel value is particularly small and large. By excluding value pixels, the effect of gradation correction can be more appropriately evaluated.

In the sixteenth aspect of the present invention,
Means for obtaining a statistical value of a pixel value of an image before correction that is an image before gradation correction and a statistical value of a pixel value of an image after correction that is an image after gradation correction;
Means for obtaining an evaluation index using the statistical value of the pixel value of the image before correction and the statistical value of the pixel value of the image after correction;
There is provided an arithmetic device characterized by comprising:
This arithmetic device functions as a device for obtaining an evaluation index.

  In the seventeenth aspect of the present invention, in the arithmetic device, a set of pixels satisfying predetermined conditions is selected as a selected pixel group from the uncorrected image and the corrected image, and pixels belonging to the selected pixel group are selected. The pixel value is acquired, and the statistical value is obtained based on the pixel value.

  Further, in an eighteenth aspect of the present invention, in the arithmetic device, there are a plurality of the selected pixel groups for each of the pre-correction image and the post-correction image, and the pixel value of each pixel belonging to each selected pixel group is acquired, The statistical value is obtained based on the pixel value.

  In the nineteenth aspect of the present invention, in the arithmetic device, the predetermined condition is a range of pixel values.

  In the twentieth aspect of the present invention, the predetermined condition is that the predetermined condition is included in a partial image that is a part of the pre-correction image and the post-correction image.

  According to a twenty-first aspect of the present invention, in the arithmetic device, the partial image is specified by an area obtained by dividing the pre-correction image and the post-correction image in a grid pattern.

In the twenty-second aspect of the present invention,
The pixels included in the pre-correction image that is the image before gradation correction are classified into a plurality of luminance value levels according to the luminance value that is the pixel value, and are included in the corrected image that is the image after gradation correction. Luminance value classification means for classifying pixels into the same luminance value level as the pixel at the same position in the pre-correction image;
Means for obtaining a statistical value of a luminance value of each pixel of a pixel group included in each luminance value level for the pre-correction image and the post-correction image;
Means for obtaining an evaluation index using the statistical value of the image before correction and the statistical value of the image after correction;
There is provided an arithmetic device characterized by comprising:

  In the twenty-third aspect of the present invention, in the arithmetic unit, the range of the pixel value is determined by a threshold value set in a portion that becomes a valley of the histogram.

In the twenty-fourth aspect of the present invention,
A partial image extracting means for extracting a pixel group included in a partial image obtained by dividing the pre-correction image, which is an image before gradation correction, and the corrected image, which is an image after gradation correction, in a grid pattern;
Means for obtaining a statistical value of a pixel value of the image before correction and a statistical value of a pixel value of the image after correction using a pixel value of each pixel of the extracted pixel group;
Means for obtaining an evaluation index using a statistical value of the pixel value of the image before correction and a statistical value of the pixel value of the image after correction;
There is provided an arithmetic device characterized by comprising:

  In the twenty-fifth aspect of the present invention, in the arithmetic device, the partial images are classified by a range of pixel values in units of pixel groups arranged in a rectangular shape.

  According to a twenty-sixth aspect of the present invention, in the arithmetic device, the partial image is determined based on a designated position on the image.

  In the twenty-seventh aspect of the present invention, the evaluation index is a ratio of a statistical value of the pixel value of the pre-correction image and a statistical value of the pixel value of the post-correction image.

  In the twenty-eighth aspect of the present invention, the statistical value is a value based on a standard deviation.

  In a twenty-ninth aspect of the present invention, the statistical value is a value based on a difference between a maximum value and a minimum value of pixel values.

  In the thirtieth aspect of the present invention, the statistical value is a value based on a difference between the maximum value and the minimum value of the specified pixel value by specifying a pixel whose pixel value is included in a predetermined ratio at the center. It is characterized by.

  According to the sixteenth to thirtieth aspects of the present invention, an apparatus that contributes to the execution of the gradation correction evaluation method according to the first to fifteenth aspects of the present invention is realized. It is possible to accurately obtain an evaluation index necessary for the gradation correction evaluation method according to the above.

  According to the present invention, it is possible to obtain an evaluation index indicating the effect of gradation correction performed on an image.

It is a figure which shows the hardware constitutions of the correction | amendment evaluation apparatus (arithmetic apparatus) which concerns on one Embodiment of this invention. It is a functional block diagram which shows the structure of the correction | amendment evaluation apparatus which concerns on the 1st Embodiment of this invention. It is a flowchart which shows the processing operation of the correction | amendment evaluation apparatus which concerns on the 1st Embodiment of this invention. It is a figure which shows the concept of the luminance value level classification | category of the pixel in the image before correction | amendment, and the image after correction | amendment. It is a figure which shows the relationship between three definitions of a range width | variety, and the histogram of the luminance value of a pixel. It is a figure which shows the mode of the change of the histogram before a correction | amendment about a certain image, and a correction | amendment, and a range width | variety. It is a figure which shows a response | compatibility of the luminance value of the image before correction | amendment by the correction method based on a different tone curve, and the image after correction | amendment. It is a figure which shows a response | compatibility of the luminance value of the image before correction | amendment, and the image after correction | amendment in the case of correct | amending based on a different tone curve for every luminance value level. It is a functional block diagram which shows the structure of the correction | amendment evaluation apparatus which concerns on the 2nd Embodiment of this invention. It is a flowchart which shows the processing operation of the correction | amendment evaluation apparatus which concerns on the 2nd Embodiment of this invention. It is a figure which shows the specific example of the partial image in the whole image. It is a functional block diagram which shows the structure of the correction | amendment evaluation apparatus which concerns on the 3rd Embodiment of this invention. It is a flowchart which shows the processing operation of the correction | amendment evaluation apparatus which concerns on the 3rd Embodiment of this invention. It is a figure which shows the concept of the luminance value level classification | category of the pixel in the image before correction | amendment in the 3rd Embodiment, and the image after correction | amendment. It is a functional block diagram which shows the structure of the correction | amendment evaluation apparatus which concerns on the 4th Embodiment of this invention. It is a flowchart which shows the processing operation of the correction | amendment evaluation apparatus which concerns on the 4th Embodiment of this invention. It is a figure which shows the example of the set partial image in 4th Embodiment. It is a figure which shows an example of the evaluation result by the correction | amendment evaluation apparatus which concerns on 4th Embodiment. It is a figure which shows the mode of the search of the area | region which sets a partial image. It is a figure which shows the mode of the search of the area | region which sets a partial image. It is a figure which shows the setting of the boundary value for the classification | category to a luminance value level. It is a figure which shows a response | compatibility of the gradation value of each color of the image before correction | amendment in the case of correcting a color tone so that blue may be emphasized when an image is yellowish. FIG. 10 is a diagram illustrating correspondence between gradation values of respective colors of an image before correction and an image after correction when color tone correction is performed so that red and green are emphasized and blue is attenuated when the image is bluish. It is a figure which shows a response | compatibility with the gradation value of each color of the image before correction | amendment in the case of carrying out color tone correction | amendment by offset.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the drawings.

<1. First Embodiment>
<1.1. Device configuration>
FIG. 1 is a hardware configuration diagram of a correction evaluation apparatus (arithmetic apparatus) according to the first embodiment of the present invention. The correction evaluation apparatus 100 can be realized by incorporating a dedicated program into a general-purpose computer. As shown in FIG. 1, a correction evaluation apparatus 100 includes a CPU (Central Processing Unit) 1, a RAM (Random Access Memory) 2 as a main memory, a hard disk and a flash for storing programs and data executed by the CPU 1. A storage device 3 such as a memory, an instruction input unit 4 such as a keyboard / mouse, a data input / output I / F (interface) 5 for data communication with an external device such as a data storage medium, and a display device such as a liquid crystal display Are connected to each other via a bus. The storage device 3 stores an uncorrected image that is an image before contrast correction and an corrected image that is an image after contrast correction. Contrast correction is one of the broad concepts of gradation correction. In the present embodiment, a case where the luminance value of a pixel is used for contrast correction will be described as an example. However, the present invention changes various pixel values other than the luminance value to perform color tone correction, gamma correction, and the like. The present invention can also be applied to various gradation corrections.

  FIG. 2 is a functional block diagram showing the configuration of the correction evaluation apparatus according to the present embodiment. In FIG. 2, 10 is an image capture means, 20 is a luminance value classification means, 30 is a range width calculation means, 40 is a range change amount calculation means, and 50 is an image storage means.

  The image capture means 10 captures the pre-correction image and post-correction image specified from the images stored in the image storage means 50. The luminance value classification unit 20 classifies each pixel of the captured image into a luminance value level set in advance according to the luminance value. In this embodiment, since the luminance value is used as the pixel value, it is classified into a luminance value level which is an example of the pixel value level. For each luminance value level of each image, the range width calculating unit 30 calculates the range width of the pixels belonging to the luminance value level. The range change amount calculation means 40 calculates the change amount of the range width between the uncorrected image and the corrected image for each luminance value level. The image storage unit 50 stores a pre-correction image that is a pre-correction image to be subjected to contrast correction, and a post-correction image that is an image after performing contrast correction on the pre-correction image.

  The image capturing unit 10, the luminance value classifying unit 20, the range width calculating unit 30, and the range change amount calculating unit 40 are realized by the CPU 1 executing a program stored in the storage device 3. The image storage unit 50 stores a pre-correction image that is image data before correction that is a target of contrast correction, and a post-correction image that is image data after performing contrast correction on the pre-correction image. And is realized by the storage device 3. As the image data, a single-color gray scale image can be used, or a color image of a plurality of colors can be used. Further, as gradations of image data, it is possible to use various gradations such as those in which each color of each pixel is expressed by 6-bit 64 gradations and those expressed by 8-bit 256 gradations. it can.

  Each component shown in FIG. 2 is actually realized by installing a dedicated program in hardware such as a computer and its peripheral devices as shown in FIG. That is, the computer executes the contents of each means according to a dedicated program. In this specification, the computer means an apparatus having an arithmetic processing unit such as a CPU and capable of data processing, and is not only a general-purpose computer such as a personal computer but also a data processing terminal such as a tablet or a so-called smartphone. Including.

  The storage device 3 shown in FIG. 1 is mounted with a dedicated program for operating the CPU 1 and causing the computer to function as a correction evaluation device. By executing this dedicated program, the CPU 1 realizes functions as the image capturing means 10, the luminance value classification means 20, the range width calculation means 30, and the range change amount calculation means 40. The storage device 3 not only functions as the image storage unit 50 but also stores various data necessary for processing as the correction evaluation device.

<1.2. Processing action>
Next, the contrast correction evaluation method according to the first embodiment of the present invention will be described together with the processing operation of the correction evaluation apparatus shown in FIGS. 1 and 2 using the flowchart shown in FIG. First, the image capture means 10 captures the pre-correction image and the post-correction image stored in the image storage means 50 (step S11). If the pre-correction image and the post-correction image are grayscale images, they are captured as they are. If the pre-correction image and the post-correction image are color images, the image capturing means 10 performs gray scale conversion or extracts only a part of the values and captures them as a gray scale image.

  For example, when the pre-correction image and the post-correction image are RGB color images, grayscale conversion is performed by performing an operation of 0.299 × R + 0.587 × G + 0.114 × B for each pixel, Only the G component having a strong weight is extracted and captured as a grayscale image. When the pre-correction image and the post-correction image are color images in YCbCr format or YIQ format, only the Y component is extracted and captured as a grayscale image. For other types of color images as well, the image capturing means 10 converts and captures a grayscale image of a format in which one pixel has one value. The value of the pixel expressing the brightness captured as a gray scale image in this way is set as the luminance value.

  Therefore, the image capturing means 10 captures the pre-correction image and the post-correction image as a grayscale image in which each pixel has a luminance value. Here, the luminance value is not limited to purely expressing the luminance, but includes a value expressing the gradation of the pixel based on some standard (such as “gray level” which is the original value of the pixel). Hereinafter, a case where the pre-correction image and the post-correction image are captured as a 6-bit 64-gradation grayscale image by the image capturing unit 10 will be described as an example.

  Next, the luminance value classifying unit 20 classifies each pixel into a plurality of luminance value levels according to the luminance value of each pixel of the captured pre-correction image and post-correction image (step S12). Specifically, each pixel of the pre-correction image is classified into one of a plurality of luminance value levels, and for the post-correction image, the pixel at the same position as the pixel of the pre-correction image belongs to the same luminance value level. Classify. As a result, classified pixels exist as selected pixel groups at each luminance value level.

  The number of levels to be classified as a plurality of luminance value levels can be set as appropriate, but in this embodiment, the levels are classified into three levels, a dark part, an intermediate part, and a bright part. In addition, although it is possible to appropriately set which luminance value range is classified into a dark part, an intermediate part, and a bright part, in this embodiment, a 6-bit 64-gradation pixel having a value of 0 to 63 The boundary values are set to 16 and 48, 0 to 15 or less (less than 16) is set to the dark part, 16 to 47 or less (less than 48) is set to the intermediate part, and 48 to 63 or less is set to the bright part.

  For the pre-correction image, each pixel is classified into one of the luminance value levels according to its luminance value. The corrected image is not classified according to the luminance value of each pixel of the corrected image, but the luminance value of the pixel of the pre-correction image at the same position (same coordinate) as each pixel of the corrected image. It is classified into the same luminance value level as the level. Since the post-correction image is an image obtained by performing contrast correction on the pre-correction image, the number of horizontal and vertical (x and y) pixels is naturally the same, and the pixel positions are also one-to-one. It corresponds. Therefore, pixels at the same position (same coordinates) can be easily identified.

  Here, the concept of the luminance value level classification of the pixels in the pre-correction image and the post-correction image is shown in FIG. In FIG. 4, the left and right large rectangles indicate the entire image area of the uncorrected image and the corrected image, respectively. Each of the three small rectangles in the pre-correction image indicates one pixel, and each of the three small rectangles in the post-correction image indicates a pixel at the same position (same coordinate) as the pixel in the pre-correction image. . Among the small rectangles, those that are filled with black indicate pixels having a luminance value of 0 to 15 in 64 gradations, and those that are shaded with diagonal lines are 16 to 47 in 64 gradations. The pixels with the luminance value are filled with white, and the pixels with the luminance value of 48 or more and 63 or less of the 64 gradations are shown.

  As shown in FIG. 4, the pixels of the image before correction are classified into a dark part, an intermediate part, and a bright part according to the luminance value. However, since the pixels of the corrected image are classified into the same luminance value level as the pixels of the uncorrected image at the same position, in the example of FIG. 4, the pixels with luminance values of 48 or more and 63 or less filled in white. Even in the case of a pixel having a brightness value of 0 to 15 that is classified as a dark part and painted in black, it may be classified as an intermediate part or a bright part.

  Next, the range width calculation means 30 calculates the range width of each classified luminance value level for each of the pre-correction image and the post-correction image (step S13). The range width indicates the width of the brightness value spread, and can be determined by a plurality of definitions based on statistical values. In this embodiment, the range width is determined by three definitions, and one of them can be used. It has become. The first definition is standard deviation. For each luminance value level, the standard deviation σ of the luminance values of the included pixels is calculated, and 2σ, which is twice the standard deviation σ, is set as the range width based on the standard deviation σ. That is, a standard deviation σ is obtained as a statistical value of luminance values, and 2σ based on the standard deviation σ is set as a range width.

  The second definition is the difference between the maximum and minimum values. For each luminance value level, the difference between the maximum value and the minimum value of the luminance values of the included pixels is calculated, and the calculated difference is set as the range width. That is, the maximum value and the minimum value of the luminance value are obtained as the statistical value of the luminance value, and the difference is set as the range width.

  The third definition is a range of the number of pixels at a predetermined ratio in the center. For each luminance value level, specify the number of pixels in the center of the luminance value of the included pixels, calculate the difference between the maximum and minimum luminance values of the specified pixel, and range the calculated difference Width. That is, the maximum value and the minimum value of the luminance values of the pixels at a predetermined ratio in the center are obtained as the statistical values of the luminance values, and the difference is set as the range width. The predetermined ratio can be set as appropriate, but is set to 90% in the present embodiment. Therefore, the range width is the difference between the maximum luminance value and the minimum luminance value among the 90% pixels excluding the pixel 5% having a small luminance value and the pixel 5% having a large luminance value. In other words, the range width is the difference between the maximum luminance value and the minimum luminance value when 45% pixels are taken from the median to the larger and smaller luminance values.

  Here, the relationship between the above three definitions of the range width and the histogram of the luminance value of the pixel is shown in FIG. 5A, 5B, and 5C, the horizontal axis indicates the luminance value, and the vertical axis indicates the number of pixels. Since the horizontal axis is 6 bits and 64 gradations, it takes a value of 0 to 63. On the vertical axis, a numerical value normalized by setting “1” when the number of pixels is the maximum is shown. The histograms in FIGS. 5A, 5B, and 5C are created from the same image, and are all the same.

  A relationship between such a histogram and a range width will be described. FIG. 5A shows an example in which 2σ, which is twice the standard deviation σ, is set as the range width. In the histogram of FIG. 5, since the average value is 21 and the standard deviation is 6, 12 is the range width.

  FIG. 5B is an example in which the difference between the maximum value and the minimum value of the luminance value of the pixel is set as the range width. In the histogram of FIG. 5, since the maximum value of the luminance value is 41 and the minimum value of the luminance value is 4, the difference 37 is the range width.

  FIG. 5C identifies 90% of pixels in the center of the luminance value of the pixel (excluding 5% from the minimum and 5% from the maximum), and the maximum and minimum luminance values of the specified pixel. The difference between the values is calculated, and the calculated difference is set as the range width. The predetermined ratio can be set as appropriate, but is set to 90% in the present embodiment. In the histogram of FIG. 5, the maximum luminance value excluding the upper 5% is 32 and the minimum luminance value excluding the lower 5% is 13, so the difference 19 is the range width.

  As shown in FIG. 5, the range width varies depending on what definition is set, but the range width set by the same definition is used for the pre-correction image and the post-correction image, A fair comparison can be made. Which definition is selected as the range width is set in advance in consideration of image characteristics and the like. In addition to the above three definitions, it is possible to set the range width using the statistical value of the luminance value.

  If the range width of each brightness value level of the pre-correction image and each brightness value level of the post-correction image is calculated in step S13, then the range change amount calculation means 40 calculates the range change amount of each brightness value level. (Step S14). Specifically, for each luminance value level, by using the range width of the pre-correction image and the range width of the post-correction image, the processing according to the following [Equation 1] is executed, thereby the range of each luminance value level. A change amount RV is calculated.

[Formula 1]
RV = (range width of image after correction) / (range width of image before correction)

  Since the range change amount RV is calculated for each luminance value level, as shown in [Formula 1], for each luminance value level, the range width of the image after correction is the range width of the image before correction. It is calculated by dividing. That is, the range change amount RV is expressed by a ratio between the range width of the corrected image and the range width of the uncorrected image. This range change amount RV is used as an evaluation index for evaluating the effect of gradation correction such as contrast correction.

  When the range change amount RV is larger than 1, it indicates that the contrast is improved, and the degree of improvement is higher as the value of the range change amount RV is larger. A range change amount RV = 1 indicates that the contrast is maintained even after correction. When the range change amount RV is smaller than 1, it indicates that the contrast has deteriorated after correction. In this case, gradation collapse may occur at that luminance value level. The range change amount RV = 0 indicates that the corrected pixel value (luminance value) is constant (saturated). This means that so-called black crushing or whiteout occurs due to the correction. If contrast correction is not performed and the pre-correction image and the post-correction image are the same, the range change amount RV = 1.

  FIG. 6 shows changes in histogram and range width before and after correction for a certain image. 6A shows a histogram before correction, and FIG. 6B shows a histogram after correction in units of luminance value levels of a dark part, an intermediate part, and a bright part. Comparing the dark part, the intermediate part, and the bright part before and after the correction, the corrected image shown in FIG. 6B is compared with the range width of the pre-correction image shown in FIG. It can be seen that the range width has increased. This indicates that the range change amount RV is larger than 1 in each of the dark part, the intermediate part, and the bright part, and the contrast is improved.

  Next, the correction process is evaluated (step S15). Specifically, the range change amount RV calculated in step S14 is confirmed. If the value of the range change amount RV is large, the correction method is highly evaluated. If the value of the range change amount RV is small, the correction is made. Give a low rating for the method. As a specific presentation method of evaluation, for example, a predetermined threshold value is set, and if it is larger than the threshold value, it is indicated as “high evaluation”, if it is smaller than the threshold value, “low evaluation”, etc. Display output is possible. In addition, the evaluation can be presented by directly outputting the value of the range change amount RV on the display unit 6.

  FIG. 7 is a diagram illustrating a correspondence between luminance values of an image before correction (indicated as “input” in the figure) and an image after correction (indicated as “output” in the figure) by a correction method based on different tone curves. 7 (a), (b), and (c), the thin line segment from the lower left end to the upper right end indicates a reference line when correction is not performed, and a thick curve connecting the upper left end and the upper right end is The tone curve defining the correction method is shown. Further, the ranges divided into three on the horizontal axis and the vertical axis indicate the dark part, the intermediate part, and the bright part from the 0 side (lower left end), respectively.

  FIG. 7A shows correction based on a tone curve for emphasizing the dark portion, and the range range of the bright portion is reduced by an amount corresponding to the increase of the range range of the dark portion. That is, the correction shown in FIG. 7A improves the contrast of the dark part. When the range change amount is calculated by the correction evaluation apparatus according to the present embodiment for the pre-correction image and the post-correction image shown in FIG. 7A, the range change amount in the dark part is> 1, and the range change amount in the intermediate part < 1. The range change amount of the bright part is <1.

  FIG. 7B shows correction based on a tone curve that emphasizes the bright portion, and the dark range range is reduced by an amount corresponding to the increase of the bright range range. That is, in the correction shown in FIG. 7B, the contrast of the bright part is improved. When the range change amount is calculated by the correction evaluation apparatus according to the present embodiment for the pre-correction image and the post-correction image shown in FIG. 7B, the range change amount in the dark portion is <1, and the range change amount in the intermediate portion < 1. The range change amount of the bright part is greater than 1.

  FIG. 7C shows correction based on a tone curve that emphasizes the intermediate portion, and the range range of the bright portion and the dark portion is reduced by an amount corresponding to the increase of the range range of the intermediate portion. That is, in the correction shown in FIG. 7C, the contrast of the intermediate portion is improved. When the range change amount is calculated by the correction evaluation apparatus according to the present embodiment for the uncorrected image and the corrected image shown in FIG. 7C, the range change amount in the dark part is <1, the range change amount in the intermediate part>. 1. The range change amount of the bright part is <1.

  FIG. 8 shows the correspondence between the luminance values of the image before correction (indicated as “input” in the figure) and the image after correction (indicated as “output” in the figure) when correction based on a tone curve that differs for each luminance value level is performed. FIG. FIG. 8A shows a case where only the dark portion is corrected by a tone curve that emphasizes the dark portion. FIG. 8B shows a case where only the bright part is corrected by a tone curve that emphasizes the bright part. FIG. 8C shows a case where only the intermediate portion is corrected by a tone curve that emphasizes the intermediate portion.

  As shown in FIG. 8, by performing correction based on different tone curves for the dark portion, intermediate portion, and bright portion of the pre-correction image, the contrast of all of the dark portion, intermediate portion, and bright portion is improved. When the range change amount is calculated by the correction evaluation apparatus according to the present embodiment for the pre-correction image and the post-correction image shown in FIG. 8, the range change amount in the dark part is> 1, the range change amount in the intermediate part> 1, The range change amount of the section is> 1.

<2. Second Embodiment>
<2.1. Device configuration>
Next, a second embodiment will be described. As in the first embodiment, the hardware configuration of the correction evaluation apparatus according to the second embodiment is as shown in FIG. 1, and the correction evaluation apparatus 101 according to the second embodiment includes: This can be realized by incorporating a dedicated program into a general-purpose computer.

  FIG. 9 is a functional block diagram showing the configuration of the correction evaluation apparatus according to the present embodiment. In FIG. 9, 11 is an image capturing means, 21 is a partial image extracting means, 31 is a range width calculating means, 41 is a range change amount calculating means, and 51 is an image storage means.

  The image capturing unit 11 captures the pre-correction image and the post-correction image specified from the images stored in the image storage unit 51. The partial image extraction unit 21 extracts a partial image that is a part of the captured image. The range width calculation unit 31 calculates the range width of the pixels belonging to the partial image for each partial image of the pre-correction image and the post-correction image. The range change amount calculation means 41 calculates the change amount of the range width between the pre-correction image and the post-correction image for each partial image. The image storage unit 51 stores a pre-correction image that is a pre-correction image to be subjected to contrast correction, and a post-correction image that is an image after performing contrast correction on the pre-correction image. As described above, contrast correction is one of the broad concepts of gradation correction. In the present embodiment, a case where the luminance value of a pixel is used for contrast correction will be described as an example. However, the present invention changes various pixel values other than the luminance value to perform color tone correction, gamma correction, and the like. The present invention can also be applied to various gradation corrections.

  As in the first embodiment, the image capturing unit 11, the partial image extracting unit 21, the range width calculating unit 31, and the range change amount calculating unit 41 execute a program stored in the storage device 3. It is realized by. Similar to the first embodiment, the image storage unit 51 is a pre-correction image that is an image data before correction that is a target for contrast correction, and image data that has been subjected to contrast correction on the pre-correction image. The storage unit stores the corrected image, and is realized by the storage device 3. As the image data, a single-color gray scale image can be used, or a color image of a plurality of colors can be used. Further, as gradations of image data, it is possible to use various gradations such as those in which each color of each pixel is expressed by 6-bit 64 gradations and those expressed by 8-bit 256 gradations. it can.

  Each component shown in FIG. 9 is actually realized by mounting a dedicated program on hardware such as a computer and its peripheral devices as shown in FIG. That is, the computer executes the contents of each means according to a dedicated program. In this specification, the computer means an apparatus having an arithmetic processing unit such as a CPU and capable of data processing, and is not only a general-purpose computer such as a personal computer but also a data processing terminal such as a tablet or a so-called smartphone. Including.

  The storage device 3 shown in FIG. 1 is mounted with a dedicated program for operating the CPU 1 and causing the computer to function as a correction evaluation device. By executing this dedicated program, the CPU 1 realizes functions as the image capturing unit 11, the partial image extracting unit 21, the range width calculating unit 31, and the range change amount calculating unit 41. The storage device 3 not only functions as the image storage unit 51 but also stores various data necessary for processing as the correction evaluation device. For example, information about which area is specified as the partial image is also stored.

<2.2. Processing action>
Next, the gradation correction evaluation method according to the second embodiment of the present invention will be described together with the processing operation of the correction evaluation apparatus shown in FIGS. 1 and 9 using the flowchart shown in FIG. First, as in the first embodiment, the image capturing unit 11 captures the pre-correction image and the post-correction image stored in the image storage unit 51 (step S21). Next, the partial image extraction unit 21 extracts a partial image for each of the captured pre-correction image and post-correction image (step S22). Specifically, an area set in advance is set as a partial image, and a pixel group belonging to each partial image is extracted as a selected pixel group.

  FIG. 11 is a diagram illustrating an example of segmentation of partial images. The example of FIG. 11 shows an example in which an image to be corrected is divided into 48 partial images of 8 × 6. In the example of FIG. 11, the partial image is specified by a region divided in a grid pattern. The number of partial images can be set in advance, but it is necessary to set the number of pre-correction images and the post-correction images so that they are the same number and the same position.

  Next, the range width calculation unit 31 calculates the range width of each partial image for each of the pre-correction image and the post-correction image (step S23). Specifically, the range width is calculated for each partial image in the same manner as performed for each luminance value level in step S13 of the first embodiment. Therefore, in the second embodiment as well, as in the first embodiment, the range width depends on one of three definitions: standard deviation, difference between maximum and minimum values, and a range of the number of pixels at a predetermined ratio close to the median. Is calculated.

  If the range width of each partial image of the pre-correction image and each partial image of the post-correction image is calculated in step S23, next, the range change amount of each partial image is calculated (step S24). Specifically, for each partial image, using the range width of the pre-correction image and the range width of the post-correction image, by executing the processing according to the above [Equation 1] used in the first embodiment, A range change amount RV of each luminance value level is calculated.

  In the second embodiment, the range change amount RV is calculated for each partial image. Therefore, as shown in [Formula 1], the range width of the corrected image is corrected for each partial image. It is calculated by dividing by the range width of the previous image. This range change amount RV is used as an evaluation index for evaluating the effect of gradation correction such as contrast correction.

  Similar to the first embodiment, when the range change amount RV is larger than 1, it indicates that the contrast is improved, and the degree of improvement is higher as the value of the range change amount RV is larger. A range change amount RV = 1 indicates that the contrast is maintained even after correction. When the range change amount RV is smaller than 1, it indicates that the contrast has deteriorated after correction. In this case, gradation collapse may occur at that luminance value level. The range change amount RV = 0 indicates that the corrected pixel value (luminance value) is constant (saturated). This means that so-called black crushing or whiteout occurs due to the correction. If contrast correction is not performed and the pre-correction image and the post-correction image are the same, the range change amount RV = 1.

  Next, the correction process is evaluated (step S25). Specifically, the range change amount RV calculated in step S24 is confirmed. If the value of the range change amount RV is large, the correction method is highly evaluated. If the value of the range change amount RV is small, the correction is made. Give a low rating for the method. The higher the evaluation for each partial image, the higher the evaluation of correction for the entire image.

<3. Third Embodiment>
<3.1. Device configuration>
Next, a third embodiment will be described. As in the first and second embodiments, the hardware configuration of the correction evaluation apparatus according to the third embodiment is as shown in FIG. 1. The correction evaluation apparatus according to the third embodiment 102 can be realized by incorporating a dedicated program into a general-purpose computer.

  FIG. 12 is a functional block diagram showing the configuration of the correction evaluation apparatus according to this embodiment. In FIG. 12, 12 is an image capture means, 22 is a partial image extraction means, 23 is a luminance value classification means, 32 is a range width calculation means, 42 is a range change amount calculation means, and 52 is an image storage means.

  The image capturing unit 12 captures the pre-correction image and the post-correction image specified from the images stored in the image storage unit 52. The partial image extraction unit 22 extracts a partial image that is a part of the captured image. The luminance value classifying unit 23 classifies the extracted partial images into luminance value levels set in advance according to the luminance value. In this embodiment, since the luminance value is used as the pixel value, it is classified into a luminance value level which is an example of the pixel value level. For each luminance value level of the pre-correction image and the post-correction image, the range width calculation unit 32 calculates the range width of the pixels belonging to the luminance value level.

  The range change amount calculation means 42 calculates the change amount of the range width between the pre-correction image and the post-correction image for each luminance value level. The image storage unit 52 stores a pre-correction image that is a pre-correction image to be subjected to contrast correction, and a post-correction image that is an image after performing contrast correction on the pre-correction image. As described above, contrast correction is one of the broad concepts of gradation correction. In the present embodiment, a case where the luminance value of a pixel is used for contrast correction will be described as an example. However, the present invention changes various pixel values other than the luminance value to perform color tone correction, gamma correction, and the like. The present invention can also be applied to various gradation corrections.

  As in the first and second embodiments, the CPU 1 stores the image capturing unit 12, the partial image extracting unit 22, the luminance value classifying unit 23, the range width calculating unit 32, and the range change amount calculating unit 42 in the storage device 3. This is realized by executing a stored program. Similar to the first and second embodiments, the image storage unit 52 includes an image before correction that is image data before correction that is a target of contrast correction, and an image after contrast correction is performed on the image before correction. A storage unit that stores a corrected image that is data, and is realized by the storage device 3. As the image data, a single-color gray scale image can be used, or a color image of a plurality of colors can be used. Further, as gradations of image data, it is possible to use various gradations such as those in which each color of each pixel is expressed by 6-bit 64 gradations and those expressed by 8-bit 256 gradations. it can.

  Each component shown in FIG. 12 is actually realized by installing a dedicated program on hardware such as a computer and its peripheral devices as shown in FIG. That is, the computer executes the contents of each means according to a dedicated program. In this specification, the computer means an apparatus having an arithmetic processing unit such as a CPU and capable of data processing, and is not only a general-purpose computer such as a personal computer but also a data processing terminal such as a tablet or a so-called smartphone. Including.

  The storage device 3 shown in FIG. 1 is mounted with a dedicated program for operating the CPU 1 and causing the computer to function as a correction evaluation device. By executing this dedicated program, the CPU 1 realizes functions as the image capturing unit 12, the partial image extracting unit 22, the luminance value classifying unit 23, the range width calculating unit 32, and the range change amount calculating unit 42. become. The storage device 3 not only functions as the image storage unit 52 but also stores various data necessary for processing as the correction evaluation device. For example, information about which area is specified as the partial image is also stored.

<3.2. Processing action>
Next, a gradation correction evaluation method according to the third embodiment of the present invention will be described together with the processing operation of the correction evaluation apparatus shown in FIGS. 1 and 12 using the flowchart shown in FIG. First, as in the first and second embodiments, the image capturing unit 12 captures the pre-correction image and the post-correction image stored in the image storage unit 52 (step S31). Next, the partial image extraction unit 22 extracts a partial image for each of the captured pre-correction image and post-correction image (step S32). Specifically, a region set in advance is set as a partial image, and pixels belonging to each partial image are extracted.

  FIG. 11 is a diagram illustrating an example of segmentation of partial images. Similar to the second embodiment, the example of FIG. 11 shows an example in which the image to be corrected is divided into 48 partial images of 8 × 6. In the example of FIG. 11, the partial image is specified by a region divided in a grid pattern. The number of partial images can be set in advance, but it is necessary to set the number of pre-correction images and the post-correction images so that they are the same number and the same position.

  Next, the luminance value classifying unit 23 classifies each partial image into a plurality of luminance value levels according to the luminance value of the partial image extracted from the pre-correction image (step S33). Specifically, the partial image of the pre-correction image is classified into one of a plurality of luminance value levels, and for the post-correction image, the partial image at the same position as the partial image of the pre-correction image belongs to the same luminance value level. Classify as follows. Here, the luminance value of the partial image can be a luminance value representative of all the pixels belonging to the partial image. For example, the average value, median value, etc. of the luminance values of all the pixels belonging to the partial image can be used. Thereby, the pixels classified into each luminance value level exist as a selected pixel group.

  The number of levels to be classified as a plurality of luminance value levels can be set as appropriate, but in this embodiment, the levels are classified into three levels, a dark part, an intermediate part, and a bright part. In addition, although it is possible to appropriately set which luminance value range is classified into a dark part, an intermediate part, and a bright part, in this embodiment, a 6-bit 64-gradation pixel having a value of 0 to 63 The boundary values are set to 16 and 48, 0 to 15 or less (less than 16) is set to the dark part, 16 to 47 or less (less than 48) is set to the intermediate part, and 48 to 63 or less is set to the bright part.

  For the pre-correction image, each partial image is classified into one of the luminance value levels according to the luminance value. The corrected image is not classified according to the luminance value of each partial image of the corrected image, but the partial image of the pre-correction image at the same position (same coordinate) as each partial image of the corrected image is classified. It is classified into the same luminance value level as the luminance value level. Since the post-correction image is an image obtained by performing contrast correction on the pre-correction image, the number of horizontal and vertical (x and y) pixels is naturally the same, and the pixel positions are also one-to-one. It corresponds. The size (pixel size) and position of the partial image are also the same. Therefore, partial images at the same position (same coordinates) can be easily specified.

  Here, the concept of the luminance value level classification of the pixels in the pre-correction image and the post-correction image is shown in FIG. In FIG. 14, the left and right large rectangles indicate the entire image area of the uncorrected image and the corrected image, respectively. In FIG. 14, three small rectangles in the pre-correction image indicate partial images selected for explanation from among the 8 × 6 48 partial images shown in FIG. 11. The three small rectangles in the corrected image indicate partial images at the same position (same coordinates) as the three partial images in the pre-correction image. Among the partial images, those filled in black indicate partial images having a luminance value of 0 or more and 15 or less in 64 gradations, and those shaded with diagonal lines are 16 or more 47 in 64 gradations. The partial images with the following luminance values are shown, and those filled with white indicate partial images with a luminance value of 48 to 63 out of 64 gradations.

  As shown in FIG. 14, the partial image of the pre-correction image is classified into a dark part, an intermediate part, and a bright part according to the luminance value. However, since the partial image of the corrected image is classified into the same luminance value level as the partial image of the pre-correction image at the same position, in the example of FIG. Even an image may be classified into a middle part and a bright part even if it is a partial image with a brightness value of 0 to 15 that is classified as a dark part and is painted black. All pixels included in the partial image classified into a certain luminance value level are classified into the luminance value level of the partial image regardless of the luminance value. For example, even if a certain pixel has a luminance value of 0 or more and 15 or less, if the luminance value level of a partial image including the pixel is a bright portion, it is classified as a bright portion. In this way, the classification into the luminance value level of each pixel is performed simultaneously with the classification of the partial images.

  Next, the range width calculation means 32 calculates the range width of each classified luminance value level for each of the pre-correction image and the post-correction image (step S34). Specifically, the range width is calculated in the same manner as performed for each luminance value level in step S13 of the first embodiment. Accordingly, in the third embodiment as well, as in the first embodiment, the range width depends on one of three definitions: standard deviation, difference between maximum and minimum values, and a range of a predetermined number of pixels close to the median. Is calculated.

  If the range width of each partial image of the pre-correction image and each partial image of the post-correction image is calculated in step S34, the range change amount of each luminance value level is then calculated (step S35). Specifically, for each luminance value level, by using the range width of the pre-correction image and the range width of the post-correction image, the process according to [Formula 1] used in the first embodiment is executed. Then, the range change amount RV of each luminance value level is calculated.

  In the third embodiment, the range change amount RV is calculated for each luminance value level. Therefore, as shown in the above [Equation 1], the range width of the corrected image is set for each luminance value level. , By dividing by the range width of the uncorrected image. That is, the range change amount RV is expressed by a ratio between the range width of the corrected image and the range width of the uncorrected image. This range change amount RV is used as an evaluation index for evaluating the effect of gradation correction such as contrast correction.

  Similar to the first and second embodiments, when the range change amount RV is larger than 1, it indicates that the contrast is improved, and the degree of improvement is higher as the value of the range change amount RV is larger. A range change amount RV = 1 indicates that the contrast is maintained even after correction. When the range change amount RV is smaller than 1, it indicates that the contrast has deteriorated after correction. In this case, gradation collapse may occur at that luminance value level. The range change amount RV = 0 indicates that the corrected pixel value (luminance value) is constant (saturated). This means that so-called black crushing or whiteout occurs due to the correction. If contrast correction is not performed and the pre-correction image and the post-correction image are the same, the range change amount RV = 1.

  Next, the correction process is evaluated (step S36). Specifically, the range change amount RV calculated in step S35 is confirmed. If the value of the range change amount RV is large, the correction method is highly evaluated. If the value of the range change amount RV is small, the correction is made. Give a low rating for the method. A specific presentation method for evaluation can be performed in the same manner as in the first and second embodiments.

  In the third embodiment, as in the first embodiment, a range change amount RV for each luminance value level is obtained for the entire image. The difference from the first embodiment is whether the pixel of the pre-correction image for determining the luminance value level is one pixel or a plurality of pixels constituting the partial image. When one pixel is used as a reference, there is a tendency that a step is generated in the histogram obtained for each luminance value level. On the other hand, when a plurality of pixels constituting the partial image are used as a reference, the histogram obtained for each luminance value level has a smooth shape.

<4. Fourth Embodiment>
<4.1. Device configuration>
Next, a fourth embodiment will be described. The hardware configuration of the correction evaluation apparatus according to the fourth embodiment is as shown in FIG. 1 as in the first to third embodiments, and the correction evaluation apparatus according to the fourth embodiment. 103 can be realized by incorporating a dedicated program into a general-purpose computer.

  FIG. 15 is a functional block diagram showing a configuration of the correction evaluation apparatus according to the present embodiment. In FIG. 15, 13 is an image capture means, 24 is a partial image setting means, 25 is a partial image extraction means, 33 is a range width calculation means, 43 is a range change amount calculation means, and 53 is an image storage means.

  The image capturing unit 13 captures the pre-correction image and the post-correction image specified from the images stored in the image storage unit 53. The partial image setting unit 24 sets a partial image to be extracted from the captured image. The partial image extraction unit 25 extracts the partial image set by the partial image setting unit 24. The range width calculation unit 33 calculates the range width of the pixels belonging to the partial image for each partial image of the pre-correction image and the post-correction image. The range change amount calculation unit 43 calculates the change amount of the range width between the pre-correction image and the post-correction image for each partial image. The image storage unit 53 stores a pre-correction image that is a pre-correction image that is a target for contrast correction, and a post-correction image that is an image after performing contrast correction on the pre-correction image. As described above, contrast correction is one of the broad concepts of gradation correction. In the present embodiment, a case where the luminance value of a pixel is used for contrast correction will be described as an example. However, the present invention changes various pixel values other than the luminance value to perform color tone correction, gamma correction, and the like. The present invention can also be applied to various gradation corrections.

  The image capturing unit 13, the partial image setting unit 24, the partial image extracting unit 25, the range width calculating unit 33, and the range change amount calculating unit 43 are stored in the storage device 3 by the CPU 1 as in the first to third embodiments. This is realized by executing a stored program. As in the first to third embodiments, the image storage unit 53 includes an image before correction that is image data before correction that is a target of contrast correction, and an image after contrast correction is performed on the image before correction. A storage unit that stores a corrected image that is data, and is realized by the storage device 3. As the image data, a single-color gray scale image can be used, or a color image of a plurality of colors can be used. Further, as gradations of image data, it is possible to use various gradations such as those in which each color of each pixel is expressed by 6-bit 64 gradations and those expressed by 8-bit 256 gradations. it can.

  Each component shown in FIG. 15 is actually realized by mounting a dedicated program on hardware such as a computer and its peripheral devices as shown in FIG. That is, the computer executes the contents of each means according to a dedicated program. In this specification, the computer means an apparatus having an arithmetic processing unit such as a CPU and capable of data processing, and is not only a general-purpose computer such as a personal computer but also a data processing terminal such as a tablet or a so-called smartphone. Including.

  The storage device 3 shown in FIG. 1 is mounted with a dedicated program for operating the CPU 1 and causing the computer to function as a correction evaluation device. By executing this dedicated program, the CPU 1 realizes functions as the image capturing unit 13, the partial image setting unit 24, the partial image extraction unit 25, the range width calculation unit 33, and the range change amount calculation unit 43. become. The storage device 3 not only functions as the image storage unit 53 but also stores various data necessary for processing as the correction evaluation device.

<4.2. Processing action>
Next, the gradation correction evaluation method according to the fourth embodiment of the present invention will be described together with the processing operation of the correction evaluation apparatus shown in FIGS. 1 and 15 using the flowchart shown in FIG. First, as in the first to third embodiments, the image capturing unit 13 captures the pre-correction image and the post-correction image stored in the image storage unit 53 (step S41). Next, the partial image setting unit 24 sets a partial image for the captured pre-correction image (step S42). Specifically, first, the user designates a rectangular area for setting a partial image using the instruction input unit 4 while viewing the pre-correction image displayed on the display unit 6. Designation of a rectangular area using the instruction input unit 4 such as a mouse can be performed by a known method. Since it is necessary to designate the same position as a partial image in the pre-correction image and the post-correction image, when a rectangular area is designated in one image, the rectangular area is designated at the same position in the other image. Then, the partial image setting unit 24 sets the designated rectangular area as a partial image. The setting of the partial image can be performed by other different methods. The setting of partial images by other different methods will be described later.

  FIG. 17 is a diagram illustrating an example of a set partial image. The example of FIG. 17 shows an example in which three partial images are set in the pre-correction image. In this case, the user has designated three areas. The example of FIG. 17 shows an example in which the three partial images have the same size (vertical and horizontal pixel numbers). However, since the size is actually the size of the area specified by the user, they are not necessarily the same. It will not be. The partial image setting unit 24 also sets a partial image at the same coordinates as the corrected image for the corrected image. Next, the partial image set by the partial image extraction means 25 is extracted from each of the captured pre-correction image and post-correction image (step S43). Specifically, an area set in advance is set as a partial image, and a pixel group belonging to each partial image is extracted as a selected pixel group.

  Next, the range width calculation means 33 calculates the range width of each partial image for each of the pre-correction image and the post-correction image (step S44). Specifically, the range width is calculated for each partial image in the same manner as executed in step S23 of the second embodiment. Therefore, in the fourth embodiment, as in the first to third embodiments, any one of the three definitions of the standard deviation, the difference between the maximum value and the minimum value, and the range of the predetermined number of pixels close to the median value is used. To calculate the range width.

  If the range width of each partial image of the pre-correction image and each partial image of the post-correction image is calculated in step S44, next, the range change amount of each partial image is calculated (step S45). Specifically, for each partial image, using the range width of the pre-correction image and the range width of the post-correction image, the process according to [Formula 1] used in the first to third embodiments is executed. Thus, the range change amount RV of each luminance value level is calculated.

  In the fourth embodiment, the range change amount RV is calculated for each partial image. Therefore, as shown in [Formula 1], the range width of the corrected image is corrected for each partial image. It is calculated by dividing by the range width of the previous image. This range change amount RV is used as an evaluation index for evaluating the effect of gradation correction such as contrast correction.

  As in the first to third embodiments, when the range change amount RV is larger than 1, it indicates that the contrast is improved, and the degree of improvement is higher as the value of the range change amount RV is larger. A range change amount RV = 1 indicates that the contrast is maintained even after correction. When the range change amount RV is smaller than 1, it indicates that the contrast has deteriorated after correction. In this case, gradation collapse may occur at that luminance value level. The range change amount RV = 0 indicates that the corrected pixel value (luminance value) is constant (saturated). This means that so-called black crushing or whiteout occurs due to the correction. If contrast correction is not performed and the pre-correction image and the post-correction image are the same, the range change amount RV = 1.

  Next, the correction process is evaluated (step S46). Specifically, the range change amount RV calculated in step S45 is confirmed. If the value of the range change amount RV is large, the correction method is highly evaluated, and if the value of the range change amount RV is small, the correction is made. Give a low rating for the method. The higher the evaluation for each partial image, the higher the evaluation of correction for the entire image.

  FIG. 18 is a diagram illustrating an example of an evaluation result by the correction evaluation apparatus according to the present embodiment. As shown in FIG. 18, an image before correction (denoted as “original image” in the figure) and two corrected images corrected by two correction methods (correction A and correction B) (“correction A” in the figure). A total of three images (denoted as “correction B”) are arranged side by side. In each image, four designated rectangular areas are displayed as described above. Each rectangular area is associated with a bright part, an intermediate part (bright), an intermediate part (dark), or a dark part, which is set by the subjectivity of the person who specified the rectangular area. Is. As shown in FIG. 18, the range change amount calculated by the process of step S <b> 45 is associated with “correction A” and “correction B” for each rectangular area at the bottom of the three images. The The evaluation result as shown in FIG. 18 may be displayed on the display unit 6 or may be printed out on a printer via the data input / output I / F 5.

<4.3. Partial image settings>
As described above, the setting of the partial image in step S42 can be performed by different methods. Hereinafter, three different methods for setting partial images will be described.

<4.3.1. Setting by searching around specified area>
First, setting by searching around the designated area will be described. Also in this case, the user designates a rectangular area using the instruction input unit 4 while viewing the pre-correction image displayed on the display unit 6 as in the above example. However, in the above example, the designated rectangular area is set as a partial image as it is, but here, the designated rectangular area is used to search for a rectangular area as a partial image.

  Specifically, the user designates a rectangular area as a designated area and designates a condition for luminance value. As the condition of the luminance value, an upper limit and a lower limit that specify the range of the luminance value are designated. Since the range of the luminance value may not be too wide, it is specified to be an appropriate range according to the purpose.

  When the designated area and the luminance value range are designated, the partial image setting unit 24 searches for a search target area which is a rectangular area as a partial image. Specifically, a search target area having a luminance value as a specified condition is searched in a predetermined range around the specified area. FIG. 19A shows the relationship between the designated area and the search target area. In FIG. 19A, the search target area having the same size (the same number of pixels) as the designated area is indicated by a broken-line rectangle. It is also possible to set the size of the search target area to a size different from the designated area. The coordinate value of the pixel is represented by (x, y) in a two-dimensional coordinate system, where the horizontal direction in the drawing is the x axis and the vertical direction in the drawing is the y axis.

  Then, a luminance histogram is calculated for each search target area, with all rectangular areas overlapping one or more pixels as the specified area as search target areas. The search order can be freely set, but in the example of FIG. 19A, as indicated by the arrow, the value of x is incremented by 1 with y fixed, and when x reaches the maximum. , Y is incremented by 1 and x is returned to the minimum, the region to be searched is moved, and luminance histograms are sequentially calculated.

  Then, the partial image setting unit 24 sets, as a partial image, a search target region for which the calculated luminance histogram satisfies a specified condition. Various criteria can be set as to whether or not the search target area is a specified condition. For example, it can be set on the basis that a predetermined value or more of pixel values are included in the range of the luminance value as the specified condition. In this case, the search target area is set as a partial image when a luminance value of a predetermined percentage or more of the number of pixels included in the search target area falls within the luminance value range of the specified condition.

  When there are a plurality of search target areas that satisfy the designated conditions, the search can be narrowed down by various methods. As a narrowing-down method, the search target area may be displayed as a candidate area on the screen and selected by the user, or the search target area with the highest degree of condition matching may be set as a partial image. Good.

<4.3.2. Setting by searching around specified points>
Next, setting by searching around the designated point will be described. In this case, the user designates one point as a designated point using the instruction input unit 4 while viewing the pre-correction image displayed on the display unit 6. Then, a rectangular area as a partial image is searched using the designated point designated.

  Specifically, the user designates one point on the pre-correction image as a designated point and designates a condition for luminance value. As the condition of the luminance value, an upper limit and a lower limit that specify the range of the luminance value are designated as in the above example. Since the range of the luminance value may not be too wide, it is specified to be an appropriate range according to the purpose.

  When the designated point and the range of the luminance value are designated, the partial image setting unit 24 searches for a rectangular area as a partial image. Specifically, a rectangular area having a luminance value as a designated condition is searched in a predetermined range around the designated point. FIG. 19B shows the relationship between the designated area and the search target area. In FIG. 19B, a search target area having a predetermined size (number of pixels) is indicated by a broken-line rectangle. The size of the search target area can be set to an arbitrary size. The coordinate value of the pixel is represented by (x, y) in a two-dimensional coordinate system, where the horizontal direction in the drawing is the x axis and the vertical direction in the drawing is the y axis.

  Then, a luminance histogram is calculated for each search target area, with all the rectangular areas within a predetermined distance range from the specified point as search target areas. As the predetermined distance range, for example, one or more pixels in the rectangular area can be set as a range within 3 pixels from the designated point. The search order can be set freely, but in the example of FIG. 19B, as indicated by the arrow, the value of x is incremented by 1 with y fixed, and when x reaches the maximum. , Y is incremented by 1 and x is returned to the minimum, the region to be searched is moved, and luminance histograms are sequentially calculated.

  Then, the partial image setting unit 24 sets, as a partial image, a search target region for which the calculated luminance histogram satisfies a specified condition. Various criteria can be set as to whether or not the search target area is a specified condition. For example, it can be set on the basis that a predetermined value or more of pixel values are included in the range of the luminance value as the specified condition. In this case, the search target area is set as a partial image when a luminance value of a predetermined percentage or more of the number of pixels included in the search target area falls within the luminance value range of the specified condition.

  When there are a plurality of search target areas that satisfy the designated conditions, the search can be narrowed down by various methods. As a narrowing-down method, the search target area may be displayed as a candidate area on the screen and selected by the user, or the search target area with the highest degree of condition matching may be set as a partial image. Good.

<4.3.3. Setting by brightness value from the whole image>
Next, setting by luminance value from the entire image will be described. In this case, the user uses the instruction input unit 4 to specify only the luminance value condition. As the condition of the luminance value, an upper limit and a lower limit that specify the range of the luminance value are designated as in the above example. Since the range of the luminance value may not be too wide, it is specified to be an appropriate range according to the purpose.

  When the range of the luminance value is designated, the partial image setting unit 24 searches for a rectangular area as a partial image. Specifically, a rectangular area having a luminance value as a specified condition is searched in the entire range of the pre-correction image. FIG. 20 shows the relationship between the pre-correction image and the search target area. In FIG. 20, a search target region having a predetermined size (number of pixels) is indicated by a broken-line rectangle. The size of the search target area can be set to an arbitrary size. The coordinate value of the pixel is represented by (x, y) in a two-dimensional coordinate system, where the horizontal direction in the drawing is the x axis and the vertical direction in the drawing is the y axis.

  Then, a luminance histogram is calculated for each search target area, with all rectangular areas within the entire range of the pre-correction image as search target areas. The search order can be set freely, but in the example of FIG. 20, as indicated by the arrow, the value of x is incremented by 1 with y fixed, and when x reaches the maximum, y is increased. The search target region is moved by performing a process of increasing x by 1 and returning x to the minimum, and sequentially calculating the luminance histogram.

  Then, the partial image setting unit 24 sets, as a partial image, a search target region for which the calculated luminance histogram satisfies a specified condition. Various criteria can be set as to whether or not the search target area is a specified condition. For example, it can be set on the basis that a predetermined value or more of pixel values are included in the range of the luminance value as the specified condition. In this case, the search target area is set as a partial image when a luminance value of a predetermined percentage or more of the number of pixels included in the search target area falls within the luminance value range of the specified condition.

  When there are a plurality of search target areas that satisfy the designated conditions, the search can be narrowed down by various methods. As a narrowing-down method, the search target area may be displayed as a candidate area on the screen and selected by the user, or the search target area with the highest degree of condition matching may be set as a partial image. Good.

<5. Classification into luminance value levels>
In the first and third embodiments, when classifying pixels into luminance value levels, a plurality of luminance value levels are set using boundary values set in advance. In the first and third embodiments, the boundary value can be determined from the histogram of the entire image before correction. Hereinafter, determination of the boundary value from the histogram of the entire image before correction will be described.

  FIG. 21 is a diagram showing a histogram obtained from an 8-bit 256-gradation image having a value of 0-255. When the boundary values are set in advance to 64 and 192, as shown in FIG. 21A, 0 to 63 (less than 64) is a dark part, 64 to 191 or less (less than 192) is an intermediate part, and 192 or more 255 or less is set as the bright portion.

  However, in the case of the histogram shown in FIG. 21, there are many biases of pixels toward the dark part, which cannot be said to be a preferable classification. Considering such a situation, the boundary value is determined from the histogram of the entire image before correction. Specifically, in the histogram expressing the frequency distribution of luminance values, the luminance value of a valley portion that is a portion having a lower frequency than the vicinity is used as the boundary value. The part to be a valley can be determined by a known method. For example, an inflection point is specified by formulating a histogram and performing differentiation twice, and the luminance value of the inflection point is determined as a boundary value. The number of boundary values can be set as appropriate. When the number of boundary values is set to four, the number of inflection points is determined to be four in descending order of frequency change before and after. Then, the luminance values are classified into five luminance value levels based on the determined boundary value. FIG. 21B shows a histogram classified into five luminance value levels by setting a valley portion as a boundary value.

  When determining the valley portion, a predetermined condition is added instead of selecting a predetermined number from the entire range of luminance values, one from 0 to 63, one from 64 to 128, etc. It may be.

<6. Modification>
In the above embodiment, the case where the contrast correction, the bright part correction, and the dark part correction are performed by changing the luminance value of the pixel has been described. However, as described above, the present invention uses various pixel values other than the luminance value. By changing it, various tone corrections such as color tone correction, brightness correction, and gamma correction can be performed. As an example, a modification example in which color tone correction is performed using the gradation value of each color of a pixel will be described. In an image taken with a camera or the like, the image may become reddish or bluish depending on the setting at the time of shooting. In that case, it is possible to obtain (correct) a well-balanced image by applying gradation correction (tone correction) for each color. The color tone correction is also applied when changing colors such as sky, sea, sunset, etc., to a color of your choice.

  22 to 24 show correspondences between gradation values of respective colors of pixels of an image before correction (indicated as “input” in the figure) and an image after correction (indicated as “output” in the figure) by a correction method based on different tone curves. FIG. The thin line segment from the lower left end to the upper right end in each of FIGS. 22 to 24 shows a reference line when correction is not performed, as in FIGS. 7 and 8, and straight lines and curves deviated from the reference line are The tone curve defining the correction method is shown. Further, the ranges divided into three on the horizontal axis and the vertical axis indicate the dark part, the intermediate part, and the bright part from the 0 side (lower left end), respectively.

FIG. 22 is a diagram illustrating the correspondence between the gradation values of the respective colors of the pre-correction image and the post-correction image when the color tone correction is performed to enhance blue when the image is yellowish. If the image is yellowish, it is sufficient to increase blue or weaken red and green. In the example of FIG. 22, blue is intensified. When the range change amount RV is calculated with the gradation value of each color,
Red Dark part RV = 1, Intermediate part RV = 1, Bright part RV = 1
Green Dark part RV = 1, Intermediate part RV = 1, Bright part RV = 1
Blue Dark part RV> 1, Intermediate part RV <1, Bright part RV <1
The result is obtained. In the example of FIG. 22, it can be seen that only the blue color is corrected, and correction is performed to widen the range of gradation values in the dark part of blue.

FIG. 23 is a diagram illustrating the correspondence between the gradation values of the respective colors of the pre-correction image and the post-correction image when the tone correction is performed so that red and green are emphasized and blue is attenuated when the image is bluish. If the image is bluish, red or green should not be changed or strengthened, and blue should not be changed or weakened. In the example of FIG. 23, red and green are strengthened and blue is weakened. When the range change amount RV is calculated with the gradation value of each color,
Red Dark part RV> 1, Intermediate part RV <1, Bright part RV <1
Green Dark part RV> 1, Intermediate part RV <1, Bright part RV <1
Blue Dark part RV <1, Intermediate part RV <1, Bright part RV> 1
The result is obtained. In the example of FIG. 23, it can be seen that red and green are corrected to increase the range of gradation values in the dark area, and blue are corrected to expand the range of gradation values in the bright area.

FIG. 24 is a diagram showing the correspondence between the gradation values of the respective colors of the pre-correction image and the post-correction image when color tone correction is performed with an offset. If the image is bluish, red and green should not be changed or strengthened, and blue should not be changed or weakened. In the example of FIG. 23, when the image is bluish, red and green are strengthened and blue is weakened by offset. When the range change amount RV is calculated with the gradation value of each color,
Red Dark part RV = 1, Intermediate part RV = 1, Bright part RV <1
Green Dark portion RV = 1, Intermediate portion RV = 1, Bright portion RV <1 (larger than red bright portion RV)
Blue Dark part RV <1, Intermediate part RV = 1, Bright part RV = 1
The result is obtained. In the example of FIG. 24, red is offset (plus direction) so as to narrow the range of the bright portion, and green is also narrowed similarly to the bright portion, but when the range change amount RV is compared, the offset (plus) is greater than that of red. (Direction) is low. In blue, since the range change amount RV of the dark part is low, it can be seen that the blue part is offset (minus direction) so as to narrow the dark part. Note that the correction of the tone curve inclination control shown in FIGS. 22 and 23 and the offset correction shown in FIG. 24 can be used in combination, and can be evaluated in the same manner.

  The preferred embodiments of the present invention have been described above. However, the present invention is not limited to the above embodiments, and various modifications can be made. For example, in the above-described embodiment, the luminance value is calculated using various values of the pixel, and the range change amount that is the evaluation index is calculated using the luminance value. As described above, the pixel value for calculating the range change amount is not limited to the luminance value, and may be any value that reflects some value of the pixel.

  In the first embodiment, the plurality of luminance value levels are classified into the three levels of the dark portion, the intermediate portion, and the bright portion. However, the luminance value levels may be classified into more luminance value levels. For example, a plurality of intermediate portions can be provided instead of only one. In the example shown in FIG. 18, two intermediate portions are provided.

  Further, the correction evaluation apparatus described in the above embodiment is not necessarily used for correction evaluation, and can be used as an arithmetic apparatus that calculates a range change amount. When used as an arithmetic device, in order to adjust the brightness of the backlight of the display unit based on the calculated range change amount by incorporating the arithmetic device into a tablet equipped with a display unit such as a liquid crystal display. Can be used.

1 ... CPU (Central Processing Unit)
2 ... RAM (Random Access Memory)
3 ... Storage device 4 ... Instruction input unit 5 ... Data input / output I / F
6 ... Display unit 10, 11, 12, 13 ... Image capturing means 20, 23 ... Luminance value classifying means 21, 22, 25 ... Partial image extracting means 24 ... Partial image setting means 30, 31, 32, 33 ... Range width calculating means (means for obtaining statistical values)
40, 41, 42, 43 ... Range change amount calculation means (means for obtaining evaluation index)
50, 51, 52, 53 ... Image storage means 100, 101, 102, 103 ... Correction evaluation apparatus (arithmetic apparatus)

Claims (31)

Obtain the statistical value of the pixel value of the pre-correction image that is the image before the gradation correction and the statistical value of the pixel value of the corrected image that is the image after the gradation correction,
Using the statistical value of the pixel value of the image before correction and the statistical value of the pixel value of the image after correction, an evaluation index is obtained,
An evaluation method for gradation correction, wherein gradation evaluation is evaluated using the evaluation index.   A set of pixels satisfying a predetermined condition is selected as a selected pixel group from the pre-correction image and the post-correction image, and pixel values of pixels belonging to the selected pixel group are obtained, and the statistics are calculated based on the pixel values. 2. The gradation correction evaluation method according to claim 1, wherein a value is obtained.   A plurality of the selected pixel groups exist for each of the pre-correction image and the post-correction image, pixel values of pixels belonging to each selected pixel group are acquired, and the statistical value is obtained based on the pixel values. The evaluation method for gradation correction according to claim 2.   4. The gradation correction evaluation method according to claim 2, wherein the predetermined condition is a range of pixel values.   4. The gradation correction evaluation method according to claim 2, wherein the predetermined condition is included in a partial image that is a part of the pre-correction image and the post-correction image.   The gradation correction evaluation method according to claim 5, wherein the partial image is specified by an area obtained by dividing the pre-correction image and the post-correction image in a grid pattern. The pixels included in the pre-correction image that is the image before gradation correction are classified into a plurality of pixel value levels according to the pixel values, and the pixels included in the corrected image that is the image after gradation correction are Classify the same pixel value level as the pixel at the same position in the pre-correction image,
For the pre-correction image and the post-correction image, obtain a statistical value of the pixel value of each pixel of the pixel group included in each pixel value level,
Using the statistical value of the pixel value of the image before correction and the statistical value of the pixel value of the image after correction, an evaluation index is obtained,
An evaluation method for gradation correction, wherein gradation evaluation is evaluated using the evaluation index.   8. The gradation correction range according to claim 2, wherein the range of pixel values is determined by a threshold value set in a portion that becomes a valley of a histogram. 9. Evaluation method. The pixel value of each pixel of the pixel group included in the partial image obtained by dividing the pre-correction image, which is an image before gradation correction, and the post-correction image, which is an image after gradation correction, into a grid pattern, before the correction Obtain the statistical value of the pixel value of the image and the statistical value of the pixel value of the corrected image,
Using the statistical value of the pixel value of the image before correction and the statistical value of the pixel value of the image after correction, an evaluation index is obtained,
An evaluation method for gradation correction, wherein gradation evaluation is evaluated using the evaluation index.   10. The gradation correction evaluation method according to claim 5, wherein the partial images are classified according to a range of pixel values in units of pixel groups arranged in a rectangular shape. 11.   The gradation correction evaluation method according to claim 5, wherein the partial image is determined based on a designated position on the image.   The floor according to any one of claims 1 to 11, wherein the evaluation index is a ratio of a statistical value of a pixel value of the pre-correction image and a statistical value of a pixel value of the post-correction image. Evaluation method for tonal correction.   13. The gradation correction evaluation method according to claim 1, wherein the statistical value is a value based on a standard deviation.   The gradation correction evaluation method according to claim 1, wherein the statistical value is a value based on a difference between a maximum value and a minimum value of pixel values.   2. The statistical value is a value based on a difference between a maximum value and a minimum value of a pixel value of a specified pixel that identifies a pixel whose pixel value is included in a predetermined ratio in the center. 13. The gradation correction evaluation method according to any one of items 12. Means for obtaining a statistical value of a pixel value of an image before correction that is an image before gradation correction and a statistical value of a pixel value of an image after correction that is an image after gradation correction;
Means for obtaining an evaluation index using the statistical value of the pixel value of the image before correction and the statistical value of the pixel value of the image after correction;
An arithmetic device comprising:   A set of pixels satisfying a predetermined condition is selected as a selected pixel group from the pre-correction image and the post-correction image, and pixel values of pixels belonging to the selected pixel group are obtained, and the statistics are calculated based on the pixel values. The arithmetic unit according to claim 16, wherein a value is obtained.   A plurality of the selected pixel groups exist for each of the pre-correction image and the post-correction image, the pixel value of each pixel belonging to each selected pixel group is acquired, and the statistical value is obtained based on the pixel value. The arithmetic unit according to claim 17.   The computing device according to claim 17 or 18, wherein the predetermined condition is a range of pixel values.   19. The arithmetic device according to claim 17, wherein the predetermined condition is included in a partial image that is a part of the pre-correction image and the post-correction image.   21. The arithmetic device according to claim 20, wherein the partial image is specified by an area obtained by dividing the pre-correction image and the post-correction image in a grid pattern. The pixels included in the pre-correction image that is the image before gradation correction are classified into a plurality of luminance value levels according to the luminance value that is the pixel value, and are included in the corrected image that is the image after gradation correction. Luminance value classification means for classifying pixels into the same luminance value level as the pixel at the same position in the pre-correction image;
Means for obtaining a statistical value of a luminance value of each pixel of a pixel group included in each luminance value level for the pre-correction image and the post-correction image;
Means for obtaining an evaluation index using the statistical value of the image before correction and the statistical value of the image after correction;
An arithmetic device comprising:   The arithmetic device according to any one of claims 17, 18, 19, and 22, wherein the range of the pixel value is determined by a threshold value set in a portion that becomes a valley of a histogram. Partial image extraction means for extracting a pixel group included in each of the rectangular partial images from a pre-correction image that is an image before gradation correction and a corrected image that is an image after gradation correction;
Means for obtaining a statistical value of a pixel value of the image before correction and a statistical value of a pixel value of the image after correction using a pixel value of each pixel of the extracted pixel group;
Means for obtaining an evaluation index using a statistical value of the pixel value of the image before correction and a statistical value of the pixel value of the image after correction;
An arithmetic device comprising:   The arithmetic device according to claim 20 or 24, wherein the partial images are classified by a range of pixel values in units of pixels arranged in a rectangular shape.   The arithmetic device according to claim 20 or 24, wherein the partial image is determined based on a designated position on the image.   The calculation according to any one of claims 16 to 26, wherein the evaluation index is a ratio of a statistical value of a pixel value of the pre-correction image and a statistical value of a pixel value of the post-correction image. apparatus.   28. The arithmetic device according to claim 16, wherein the statistical value is a value based on a standard deviation.   The arithmetic unit according to any one of claims 16 to 27, wherein the statistical value is a value based on a difference between a maximum value and a minimum value of pixel values.   17. The statistical value is a value based on a difference between a maximum value and a minimum value of a pixel value that specifies a pixel whose pixel value is included in a predetermined ratio in the center. The arithmetic device according to any one of 27.   A program for causing a computer to function as the arithmetic device according to any one of claims 16 to 30.