JPH1166306A - Image quality evaluation method, apparatus and recording medium - Google Patents

Abstract

(57) [Problem] To perform more accurate image evaluation by using a contrast sensitivity characteristic of a region exceeding a perception threshold as a human visual characteristic. An image to be measured is input by an image characteristic distribution measuring unit, and converted into image density distribution information by an image information converting unit. The spatial conversion means 3 converts the density distribution information into spatial frequency data. The storage means 5 stores the spatial frequency characteristics of the human visual sense with respect to the physical characteristic amounts in the region above the human perception threshold. The spatial frequency analyzing means 4 performs a spatial frequency analysis such as performing integral calculation after obtaining the product of the visual spatial frequency characteristic in the storage means 5 and the spatial frequency data converted by the spatial converting means 3. The evaluation result is output from the evaluation result output means 6.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a method, an apparatus, and a recording medium for evaluating the quality of an image output by an image output device or the like.

[0002]

2. Description of the Related Art When evaluating image quality, there are a psychological evaluation method for quantifying the degree of image quality perceived by a human and a physical evaluation method for evaluating physical characteristics of an image itself using a measuring instrument. Psychological evaluation methods are widely used for final inspection of products, but have disadvantages such as different inspectors and changes in inspection results due to fatigue of the inspectors.

Conventionally, as an apparatus for inspecting the quality of an image,
For example, JP-A-5-284260, JP-A-1-286
There is an image evaluation method and apparatus described in Japanese Patent Publication No. 084. This image evaluation apparatus converts image information including two-dimensional position information and optical information into color information, converts the converted two-dimensional information into two-dimensional spatial frequency information by frequency analysis, After converting the spatial frequency information into one dimension,
The correction corresponding to the spatial frequency characteristic of human vision is performed, and it is possible to detect not only density (brightness) information but also saturation information and hue information as color information.

[0004]

In the apparatus and method described above, the contrast sensitivity characteristic of a human is used in order to take into account the visual characteristics of a human.

However, the characteristic curve generally used as the contrast sensitivity characteristic is a characteristic curve obtained by an experiment for obtaining a threshold value of the contrast sensitivity with respect to the spatial frequency of a person. For this reason, it has no effect on a region that exceeds the threshold value of human contrast.

That is, since the image area actually evaluated by the image evaluation apparatus or method is an area that exceeds the threshold value for perception of human contrast, the conventional image evaluation apparatus or method described There was a problem that could not be evaluated.

An object of the present invention is to provide a human visual characteristic,
An object of the present invention is to provide a more accurate image evaluation method, apparatus, and recording medium by using the contrast sensitivity characteristics of an area exceeding a perceptual threshold.

[0008]

In order to achieve the above object, according to the first aspect of the present invention, a spatial distribution of a characteristic amount of an image to be measured is measured, and distribution data of the measured characteristic amount is calculated.
It is a method of evaluating the quality of the image to be measured by converting to spatial frequency distribution data and performing frequency analysis of the converted distribution data, wherein the frequency analysis is performed above a human perception threshold. The spatial frequency characteristic of human vision for the characteristic amount in the region is referred to.

According to the second aspect of the present invention, the means for measuring the spatial distribution of the characteristic amount of the image to be measured, the means for converting the measured distribution data of the characteristic amount to the spatial frequency distribution data, An image quality evaluation device comprising means for performing frequency analysis of the distribution data obtained, wherein in a region above a human perception threshold, means for storing a spatial frequency characteristic of human vision for the characteristic amount, and the frequency When the analysis is performed, a means for referring to the storage means is provided.

According to a third aspect of the present invention, the measured characteristic amount is at least one of metric lightness, metric saturation, and metric hue of the image.

The invention according to claim 4 is characterized in that a means for selecting the characteristic quantity to be measured from a plurality of types is provided.

According to a fifth aspect of the present invention, the measuring means measures the characteristic amount in a one-dimensional direction.

According to a sixth aspect of the present invention, the measuring means measures the characteristic amount in a two-dimensional direction.

According to a seventh aspect of the present invention, a means for selecting a dimension of the distribution of the characteristic quantity to be measured is provided.

According to an eighth aspect of the present invention, the storage means is constituted by a table storing the data of the characteristic amount and the data of the spatial frequency characteristic of the human visual sense corresponding thereto.

According to a ninth aspect of the present invention, the reference means obtains the characteristic amount data not stored in the storage means by interpolating the spatial frequency characteristic data.

According to a tenth aspect of the present invention, the storage means stores, as a function, a spatial frequency characteristic of human visual perception of the characteristic amount in a region above the human perception threshold.

According to the eleventh aspect, a function of measuring a spatial distribution of a characteristic amount of an image to be measured, a function of converting the distribution data of the measured characteristic amount into a distribution data of a spatial frequency, By performing a frequency analysis of the distribution data obtained, in order to evaluate the quality of the measured image, when performing the frequency analysis, in a region above a human perception threshold, the human visual space for the characteristic amount It is a computer-readable recording medium that records a program for causing a computer to implement a function of referring to frequency characteristics.

[0019]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be specifically described below with reference to the drawings. <Embodiment 1> FIG. 1 shows the structure of Embodiment 1 of the present invention.
In the figure, 1 is an image characteristic distribution measuring means, 2 is an image information converting means, 3 is a spatial converting means, 4 is a spatial frequency analyzing means, 5 is a spatial frequency characteristic of a human visual system in an area above a perceptual threshold. Means 6 are evaluation result output means.

An image characteristic distribution measuring means 1 comprising an area sensor, a CCD camera and the like inputs a measured image. The input image information (for example, RGB signals) is converted by the image information conversion means 2 into image density distribution information (for example, L *, a *, b *) which is one of the physical characteristic amounts of the image. Is done. Next, the spatial conversion means 3 converts the density distribution information of the image into spatial frequency data by performing a Fourier transform.

The storage means 5 stores the spatial frequency characteristics of the human visual sense with respect to the physical characteristic amounts in the region above the human perception threshold, which are obtained in advance by experiments.

Equation (1) shows an example of an approximate equation of the human contrast sensitivity characteristic in a region above the perception threshold.

B (L * ＾, deltaL *, f) = (P1 · deltaL * + P2) · exp (P3 · L * ＾) · exp (P4 · f) · (1-exp ((P5 · L * ＾) + P6) · f)) + P7 (1) where L * ＾ is the average brightness, deltaL * is the maximum brightness difference, and f is the spatial frequency (cycle / degr) of the brightness variation.
ee).

The left side B of the above equation is the amount of human perception (the amount of psychology)
The right side approximates it with physical quantities. Even if a physical quantity other than the above-mentioned physical quantity (for example, retinal illuminance) is used,
Although the shape is different from that of the expression (1), it can be approximated similarly.

The above-mentioned parameters P1 to P7
Is as follows.

P1 = 2.372 P2 = 0.331 P3 = 0.144 P4 = −0.208 P5 = 0.00176 P6 = −0.176 P7 = −0.0133.

The spatial frequency analyzing means 4 calculates the product of the visual spatial frequency characteristic data in the storage means 5 and the spatial frequency data converted by the spatial converting means 3 and then performs integral calculation to obtain a constant. Perform spatial frequency analysis such as doubling. Then, although not shown, the evaluation result is output from the evaluation result output means 6 after being converted to an appropriate value by performing a predetermined operation for matching the image with the subjective evaluation result.

Although the area sensor and the CCD camera are exemplified as the image characteristic distribution measuring means in the above-described embodiment, a photomultiplier or a line sensor may be used.

<Embodiment 2> If the contrast sensitivity characteristic function in a region exceeding the perception threshold as a human visual characteristic includes a term that is not linear with respect to human perception, the function becomes complicated and handling becomes extremely difficult. This is inconvenient, and the calculation requires a lot of time.

Therefore, in the present embodiment, the calculation time is reduced by using metric lightness, metric saturation, and metric hue which are considered to be relatively linear to human senses. The second embodiment is the same as the first embodiment except that the measured physical characteristic amount is at least one of the metric lightness, the metric saturation, and the metric hue of the image.

<Embodiment 3> If the physical quantity to be measured is limited to one type as in the above-described embodiment, versatility is lacking. FIG.
Shows the configuration of the third embodiment of the present invention. In this embodiment, the physical characteristic amount selecting means 7 is provided so that the user can select in advance what kind of physical characteristic amount of an image is to be measured. When the selection unit 7 selects the brightness, the measured characteristic amount is converted into the brightness by the image information conversion unit 3. Hereinafter, the selected characteristic amount is processed in the same manner as in the first embodiment.

<Embodiment 4> When a two-dimensional spatial frequency analysis is performed to evaluate a one-dimensional image quality factor such as banding, the calculation speed becomes slow. Therefore, in the fourth embodiment, the operation speed is improved by limiting the spatial frequency operation to one dimension.

FIG. 3 shows the configuration of the fourth embodiment of the present invention.
Each of the image characteristic distribution measuring means 1 and the space transforming means 3 targets one dimension. The measured image characteristic amount is
It is assumed that the metric lightness is selected by the physical characteristic amount selecting unit 7 and is selected here. The image characteristic distribution measuring means 1 such as a photomultiplier tube performs one-dimensional scanning, and the image information converting means 2 converts the measured image characteristic amount into the one-dimensional metric brightness distribution selected by the physical characteristic amount selecting means 7. Convert to

The one-dimensional metric brightness distribution is converted into one-dimensional spatial frequency data by the space conversion means 3. Further, the spatial frequency analysis means 4 corresponding to only one dimension stores the spatial frequency of the human visual perception relative to the physical characteristic amount in an area above the human perception threshold, which is obtained in advance and stored in the storage means 5. Using the characteristics, for example, after obtaining the product of the two, one-dimensional spatial frequency analysis such as performing integral calculation and multiplying by a constant is performed. Then, although not shown, a predetermined operation for matching the image with the subjective evaluation result is performed, and after being converted into an appropriate value, the evaluation result is output from the evaluation result output means 6.

<Embodiment 5> FIG. 4 shows the configuration of Embodiment 5 of the present invention. In this embodiment, the image characteristic distribution measuring means 1
Is a two-dimensional scan, and the space conversion means 3 is a space conversion for converting the two-dimensional data into one-dimensional data.

The image characteristic quantity to be measured is selected by the physical characteristic quantity selecting means 7, and here, the metric lightness is selected. An image characteristic distribution measuring unit 1 such as an area sensor performs two-dimensional scanning, and the measured image characteristic amount is converted into a two-dimensional metric brightness distribution by the image information converting unit 2.

The two-dimensional metric lightness distribution is converted into one-dimensional spatial frequency data by the spatial conversion means 3. Further, the spatial frequency analysis means 4 corresponding to only one dimension stores the spatial frequency of the human visual perception relative to the physical characteristic amount in an area above the human perception threshold, which is obtained in advance and stored in the storage means 5. Using the characteristics, for example, after obtaining the product of the two, one-dimensional spatial frequency analysis such as performing integral calculation and multiplying by a constant is performed. Then, although not shown, a predetermined operation for matching the image with the subjective evaluation result is performed, and after being converted into an appropriate value, the evaluation result is output from the evaluation result output means 6.

Embodiment 6 When performing an evaluation corresponding to a two-dimensional image quality factor such as graininess of an image, it is necessary to perform a two-dimensional spatial frequency analysis. Since the spatial frequency calculation is performed, it is possible to cope with two-dimensional image quality factors.

FIG. 5 shows the configuration of the sixth embodiment of the present invention.
The image characteristic distribution measuring means 1 and the space transforming means 3 are both two-dimensionally configured.

The image characteristic quantity to be measured is selected by the physical characteristic quantity selecting means 7, and here, it is assumed that the metric lightness is selected. Image characteristic distribution measuring means 1 such as a photomultiplier tube performs two-dimensional scanning, and a two-dimensional metric brightness distribution is input as an image characteristic amount to be measured.
Alternatively, it is a two-dimensional metric lightness distribution input two-dimensionally by an area sensor or the like. Except that the spatial frequency analysis is performed in two dimensions, it is the same as in each of the above-described embodiments, and a description thereof is omitted.

<Embodiment 7> When the number of corresponding dimensions is one, it is not possible to cope with various image quality factors and lack versatility. Therefore, in the present embodiment, the measurement dimension can be selected. FIG. 6 shows Embodiment 7 of the present invention.
Is shown. In this embodiment, the user can specify the dimension of the physical characteristic quantity to be measured in advance by using the measurement dimension selecting means 8. Except for the dimension selection, this embodiment is the same as the above-described embodiments, and a description thereof will be omitted.

<Embodiment 8> When the spatial frequency characteristics to be used are in the form of a function, the time required for the arithmetic processing becomes longer. Therefore, in the present embodiment, the storage unit 5 is configured by a lookup table that stores characteristic amount data and corresponding data of human visual spatial frequency characteristics as grid point data. FIG. 7 shows Embodiment 8 of the present invention.
Is shown.

<Embodiment 9> If only the spatial frequency characteristics that can be used are grid point data, the accuracy of the calculation is reduced. Therefore, in the present embodiment, the data interpolation means 9 is provided in the spatial frequency analysis means 4, and when the data of the spatial frequency characteristic of the human visual sense corresponding to the spatially transformed image information data is located between the lattice points, It is obtained by interpolating the grid point data in the up table. FIG. 8 shows a configuration of the ninth embodiment of the present invention.

<Embodiment 10> When spatial frequency characteristics are obtained by interpolating grid point data, an error occurs due to interpolation. In the present embodiment, the spatial frequency characteristics of human visual perception are represented by a function of the average brightness, the maximum brightness difference, and the spatial frequency, and stored in the storage unit 5, for example, as in the above equation (1). It is. FIG. 9 shows a configuration of the tenth embodiment of the present invention.

<Embodiment 11> The present invention is not limited to the above-described embodiment, and can be realized by software. When the present invention is implemented by software, as shown in FIG. 10, a general-purpose processing device including a CPU, a ROM, a RAM, a display device, a hard disk, a keyboard, a CD-ROM drive, and a scanner is prepared. On a computer-readable recording medium such as a ROM, a program for realizing the image quality evaluation processing of the present invention is recorded. The measured image input from a CCD camera, a line sensor, or the like is temporarily stored in a hard disk or the like. When the program is started, the temporarily stored image data is read, the image is evaluated, and the evaluation result is output to an image output device such as a display or a printer.

[0046]

As described above, claims 1 and 2,
According to the eleventh aspect, when the frequency analysis of the evaluation target image is performed, since the contrast sensitivity characteristic of the region exceeding the perception threshold is used as the human visual characteristic, the accuracy of the image quality evaluation is further improved.

According to the third aspect of the present invention, at least one of metric lightness, metric saturation, and metric hue, which are considered to be relatively linear to human senses, is used. Can be shortened.

According to the fourth aspect of the present invention, since the physical quantity to be measured can be selected, the versatility is improved.

According to the fifth aspect of the present invention, since the spatial frequency calculation is limited to one dimension, the calculation speed can be improved.

According to the sixth aspect of the present invention, since the two-dimensional spatial frequency calculation is performed, the quality of the image corresponding to the two-dimensional image quality factor can be evaluated.

According to the seventh aspect of the present invention, since the measurement dimension can be selected, it is possible to cope with various image quality factors, and the versatility is further improved.

According to the eighth aspect of the present invention, the means for storing the spatial frequency characteristic of human visual perception for the characteristic amount in the region above the human perception threshold is constituted by a look-up table. The processing speed for frequency analysis is improved.

According to the ninth aspect of the present invention, when the spatial frequency characteristic data corresponding to the characteristic amount data is not in the look-up table, the data is obtained by interpolating the grid point data. improves.

According to the tenth aspect of the present invention, the means for storing the spatial frequency characteristics of the human visual perception of the characteristic amount in the region above the human perception threshold is stored as a function. Accuracy is further improved.

[Brief description of the drawings]

FIG. 1 shows a configuration of a first exemplary embodiment of the present invention.

FIG. 2 shows a configuration of a third embodiment of the present invention.

FIG. 3 shows a configuration of a fourth embodiment of the present invention.

FIG. 4 shows a configuration of a fifth embodiment of the present invention.

FIG. 5 shows a configuration of a sixth embodiment of the present invention.

FIG. 6 shows a configuration of a seventh embodiment of the present invention.

FIG. 7 shows a configuration of an eighth embodiment of the present invention.

FIG. 8 shows a configuration of a ninth embodiment of the present invention.

FIG. 9 shows a configuration of a tenth embodiment of the present invention.

FIG. 10 shows a configuration example when the present invention is realized by software.

[Explanation of symbols]

 1 image characteristic distribution measuring means 2 image information converting means 3 spatial converting means 4 spatial frequency analyzing means 5 spatial frequency characteristic storing means 6 evaluation result outputting means

Claims (11)

[Claims] 1. A method of measuring a spatial distribution of a characteristic amount of an image to be measured, converting the measured distribution data of the characteristic amount into spatial frequency distribution data, and performing a frequency analysis of the converted distribution data. According to the method for evaluating the quality of the image to be measured, when performing the frequency analysis, in a region above a human perception threshold, refer to the spatial frequency characteristics of human vision for the characteristic amount. Image quality evaluation method. 2. A means for measuring a spatial distribution of a characteristic quantity of an image to be measured, a means for converting the distribution data of the measured characteristic quantity into a spatial frequency distribution data, and a frequency of the converted distribution data. An image quality evaluation apparatus comprising means for performing analysis, in a region above a human perception threshold, means for storing a spatial frequency characteristic of human vision for the characteristic amount, and when performing the frequency analysis, An image quality evaluation device comprising means for referring to storage means. 3. The image quality evaluation device according to claim 2, wherein the measured characteristic amount is at least one of metric lightness, metric saturation, and metric hue of the image. 4. The image quality evaluation device according to claim 2, further comprising means for selecting the characteristic amount to be measured from a plurality of types. 5. The image quality evaluation apparatus according to claim 2, wherein said measuring means measures said characteristic amount in a one-dimensional direction. 6. The image quality evaluation apparatus according to claim 2, wherein said measuring means measures said characteristic amount in a two-dimensional direction. 7. The image quality evaluation apparatus according to claim 2, further comprising means for selecting a dimension of a distribution of the characteristic quantity to be measured. 8. The image quality evaluation apparatus according to claim 2, wherein said storage means comprises a table storing the data of the characteristic amount and the data of the spatial frequency characteristic of human vision corresponding to the characteristic amount data. . 9. The image according to claim 2, wherein the reference unit obtains the characteristic amount data not stored in the storage unit by interpolating the spatial frequency characteristic data. Quality evaluation device. 10. The image quality evaluation according to claim 2, wherein said storage means stores, as a function, a spatial frequency characteristic of human vision with respect to the characteristic amount in a region above the human perception threshold. apparatus. 11. A function of measuring a spatial distribution of a characteristic amount of an image to be measured, a function of converting the distribution data of the measured characteristic amount into a distribution data of a spatial frequency, and a frequency of the converted distribution data. By performing analysis, in order to evaluate the quality of the image to be measured, when performing the frequency analysis, in a region above a human perception threshold, a function of referring to a spatial frequency characteristic of human vision with respect to the characteristic amount. Computer-readable recording medium on which a program for causing a computer to realize the above is recorded.