JP2001078027A - Digital image image evaluation device - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To evaluate roughness (granularity) of a visually perceived image in an image evaluation device such as a digital hard copy by a simple calculation method. SOLUTION: An image to be evaluated is input as one-dimensional image information including an optical density level and one-dimensional position information,
Performing a discrete one-dimensional Fourier transform on the image information,
By calculating the Wiener spectrum of the Fourier transform, information indicating the spatial frequency distribution of the one-dimensional image density data is generated, and the Wiener spectrum for the spatial frequency at which the sensitivity of the spatial frequency characteristic of the human visual system is maximized. The value of the spectrum is multiplied by the standard deviation of the image density data to correct the spatial frequency, and the optically enlarging means and the illuminating means are used to compare and inspect the known standard image and the image to be evaluated. The accuracy of the evaluation value is simply compared and evaluated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital image evaluation apparatus for evaluating image noise (granularity) of a digital hard copy or the like.

[0002]

2. Description of the Related Art As image noise evaluation methods, there are a psychological evaluation for quantifying the degree of roughness of a visually perceived image and a physical evaluation for evaluating the properties of an image structure by an objectively measured amount. As a scale for physically expressing image noise in the physical evaluation, there are RMS granularity using a standard deviation of density change, a Wiener spectrum obtained by Fourier-transforming a change in density, and the like.

[0003] There is also an evaluation method in which psychological evaluation and physical evaluation are combined.
lid area) psychological graininess for images
There is an algorithm of Shaw & Dooley that predicts the に よ り from the Wiener spectrum and the average density measurement. The Wiener spectrum WS (f) is a set average of the square values of the Fourier spectrum obtained by Fourier-transforming the density fluctuation ΔD (x) from the average density obtained by scanning the image with a microdensitometer. And Dory's algorithm.

[0004]

The above-mentioned conventional algorithm is useful for evaluating an analog image having no repetitive pattern. However, a pseudo halftone image representing a halftone by the ratio of black dots in the repetitive pattern is useful. That is, when applied to a digital image, as shown in FIG. 7, psychological granularity is determined by the angle at which the image to be measured (141 lines / inch, 55% halftone dot image in this example) is scanned by the optical sensor. Cannot be used for evaluating digital images because the predicted value of the image significantly changes.

Further, since the spatial frequency analysis of the image density is performed using the two-dimensional Wiener spectrum, it takes time to analyze the data. Also, when the image density distribution of this digital image is converted into a Wiener spectrum and analyzed, the spatial frequency component of this digital image originally changes with the scanning angle, and the psychological granularity that actually perceives the visual sense Could not be sufficiently expressed.

Accordingly, the present invention provides a method of measuring a digital image with a Macbeth densitometer or the like, which makes it possible for a region having a certain image density to have a visually perceived roughness of an image.
That is, it is an object of the present invention to realize a digital image image evaluation apparatus capable of correctly evaluating the image quality as a granular feeling.

[0007]

According to a first aspect of the present invention, there is provided an image evaluation apparatus for digital images, comprising: Evaluate the granularity of any halftone digital image by combining the psychological evaluation that expresses the degree of roughness felt by the user with the physical evaluation that evaluates the degree of image roughness using objectively measured numerical values. In an image evaluation device for digital images,
The result of the physical evaluation is close to the result of the psychological evaluation.

According to a second aspect of the present invention, there is provided a digital image image evaluation apparatus according to the first aspect of the present invention, wherein the digital image image evaluation apparatus is configured to perform discrete image processing based on image density data taken from a halftone digital image. A one-dimensional Fourier transform is performed to obtain a Wiener spectrum of the digital image density data, and a quantitative evaluation is performed from the two parameters of the Wiener spectrum and the variation of the density data.

According to a third aspect of the present invention, in the digital image image evaluating apparatus according to the first aspect of the present invention, the image to be evaluated is defined by an optical density level and a one-dimensional image. Image information input means for inputting as one-dimensional image information including position information; image information storage means for storing the one-dimensional image information input by the one-dimensional image information input means; Preprocessing means for performing an original Fourier transform, and a Wiener spectrum for generating information indicating a spatial frequency distribution of one-dimensional image density data by calculating a Wiener spectrum of a Fourier transform which is an output of the preprocessing means. Calculating means for the spatial frequency at which the sensitivity of the spatial frequency characteristic of the human visual system is maximized with respect to the output of the Wiener spectrum calculating means. A spatial frequency characteristic correcting means for multiplying the value of the inner spectrum by a standard deviation of the image density data to correct a spatial frequency; an optical enlarging means and an illuminating means; And an image evaluation value inspection means for comparing and inspecting the standard image with the image to be evaluated and simply comparing and evaluating the accuracy of the image evaluation value.

[0010]

(Embodiment) Embodiment 1 of the present invention.
An embodiment of a digital image image evaluation apparatus according to claim 3 will be described below with reference to FIGS. 1 to 6. Hereinafter, the configuration of the digital image image evaluation apparatus according to the present embodiment will be described with reference to FIGS. FIG. 1 is a block diagram showing a schematic configuration of a digital image image evaluation apparatus according to the present embodiment, and FIG. 2 shows details of a configuration of an image evaluation calculation processing unit 13.

In FIG. 1, reference numeral 1 denotes a personal computer for forming a two-dimensional image which has been subjected to image processing for producing a pseudo-halftone, and 2 denotes an image evaluation sample 3 obtained by hard-copying the image processed by the personal computer 1. A microdensitometer 4 for measuring the image density of the image evaluation sample 3; Reference numeral 8 denotes a micro-spot irradiating means for irradiating the image. Reference numeral 8 denotes a micro-density value 11 by an image density output unit 10 of the micro-densitometer 4 and information 12 of a one-dimensional positional relationship by an image signal reflectance detecting unit 9 of the micro-densitometer 4. Is input to the image evaluation operation processing unit 13. The image information input unit 13 performs a series of operations on the micro-density value 11 to represent the physical quantity of the granularity of the image. An image evaluation processing unit for obtaining image evaluation value (graininess), 16 is an arithmetic result output unit for outputting a result of image evaluation processing unit 13.

In FIG. 2, reference numeral 131 denotes one-dimensional image information (micro density value 1) input by the image information input means 8.
Reference numeral 1 denotes an image density memory for storing information of one-dimensional positional relationship 12) and intermediate information as a result of performing a later-described arithmetic processing on the one-dimensional image information. Preprocessing means for performing a discrete one-dimensional Fourier transform on the information;
2 is a Wiener spectrum calculating means for calculating information representing the spatial frequency distribution of the one-dimensional image density data by calculating the Fourier transform Wiener spectrum which is the output of 2. Spatial frequency characteristic correction means for correcting the spatial frequency by multiplying the output of the image by the Wiener spectrum value corresponding to the spatial frequency at which the sensitivity of the spatial frequency characteristic of the human visual system is maximized, and the standard deviation of the image density data And
135 is provided with optical magnifying means (not shown) and illuminating means (not shown), and compares and inspects a known standard image sample and the image evaluation sample 3 which the spatial frequency characteristic correcting means 134 has, Image evaluation value inspection means for simply comparing and evaluating the accuracy of the image evaluation value, reference numeral 14 denotes a calculation processing procedure storage unit storing a series of processing procedures of the image evaluation calculation processing unit 13, and reference numeral 15 denotes a calculation processing storage unit. An arithmetic control unit that controls the image evaluation arithmetic processing unit 13 according to the procedure stored in the processing procedure storage unit 14.

A general flow of the digital image evaluation apparatus according to the present embodiment will be described below with reference to FIG. In the present embodiment, the image to be evaluated needs to be input as two-dimensional information subjected to a certain image processing. In the present embodiment, the personal computer 1 is used as two-dimensional image forming means. In the personal computer 1, image processing is performed by an organized dither method or the like to arbitrarily produce a pseudo halftone (halftone).

This image is hard-copied and output as an image evaluation sample 3 by an arbitrary printing means 2. When the Macbeth densitometer on the image evaluation sample 3 is used, the macro image density is about 1.0. Using the microdensitometer 4 for the area, the image reading area 7 is set on the spot 6 set to an arbitrary size by the minute spot irradiation means 5.
The size of this spot 6 corresponds to the visual resolution □ 50
Set to μm.

In this manner, an arbitrary image density (approximately 1.
For the region having (0), the number of data of 2 n (for example, 1024 data or 2048 data) is fetched, and the micro density value 11 of these images by the micro densitometer 4 is obtained from the image density output unit 10.

At this time, the image density memory 131 as a place for storing the micro density value 11, which is the data taken in as the image density, stores the image signal of the micro densitometer 4 so that the position where the micro density value 11 is measured can be known. A signal indicating the reflectance at each sampling point of the evaluated image input from the reflectance detection unit 9 is also stored in a form in which the one-dimensional positional relationship information 12 is held. The image density memory 13
Numeral 1 also stores an intermediate result of the calculation of the image by the image evaluation calculation processing unit 13.

The image evaluation calculation processing unit 13 is a micro-density value 1 of the image to be evaluated stored in the image density memory 131.
A series of calculations shown in FIG. 2, which will be described in detail later, are performed on the image No. 1 to obtain an image evaluation value indicating the physical quantity of the granularity of the image. The calculation processing result by the image evaluation calculation processing unit 13 is output to a calculation result output unit 16 such as a CRT display device.

Hereinafter, the image evaluation operation processing unit 13 which is a main part of the present embodiment will be described in detail with reference to FIGS. First, the preprocessing unit 132 and the Wiener spectrum calculation unit 133 will be described with reference to FIG. Micro spot 6 of micro densitometer 4
Image density data using the image density memory 1
The pre-processing unit 132 performs a one-dimensional discrete Fourier transform on the image density information of the image evaluation sample 3 stored in the storage unit 31, and then calculates a Wiener spectrum calculation unit 133
Performs the Wiener spectrum calculation processing. The one-dimensional Fourier transform F (u) of the discretely sampled image density data D is calculated inside the personal computer 1 by using a generally known method.

The Wiener spectrum P (u) is calculated by (Equation 1) using the Fourier transform F (u). P (u) = | F (u) | × | F (u) | (Equation 1) This value represents the strength of the spatial frequency F (u).

Next, the spatial frequency characteristic correction means 134 will be described with reference to FIGS. FIG. 3 is a diagram illustrating a visual spatial frequency characteristic, and FIG. 4 is a diagram illustrating an example of a power spectrum of image density data.

In order to obtain a psychophysical index of granularity of a digital image, processing is performed in consideration of visual characteristics using the Wiener spectrum P (u), that is, human visual and spatial as shown in FIG. Approximately 1 line / mm, which is the spatial frequency at which the sight shows the greatest sensitivity in terms of frequency
Is calculated for the intensity of the Wiener spectrum P (u).

As shown in FIG. 4, the spectral intensity P (u) of 1 line / mm is read from the spatial frequency characteristic of the Wiener spectrum obtained by arithmetically processing the micro density data 82 of the image evaluation sample 3. By storing the values, the calculation of the evaluation parameter value from the Wiener spectrum and the calculation of the physical quantity as the standard deviation σ of the image density are performed.

That is, the one-dimensional Wiener spectrum P (u) representing the spatial frequency characteristic of the image obtained by the above calculation, the standard deviation σ of the image density calculated on the same image evaluation sample 3, and To calculate an image evaluation value that is an index of image roughness (see (Equation 2)). Image evaluation value (granularity) = P (u) × σ (Equation 2)

As described above, the image evaluation calculation processing unit 13 for obtaining the graininess evaluation value has been described in detail. However, the calculation method used for each calculation unit is not limited as long as an equivalent result can be obtained. Obviously, a mathematical expression or an approximate expression may be used.

A series of arithmetic processing procedures are stored in the arithmetic processing procedure storage unit 14, and the arithmetic control unit 15 controls the image evaluation arithmetic processing unit 13 according to the stored procedures. Various arithmetic processing procedures stored in the arithmetic processing procedure storage unit 14 are design items that can be arbitrarily configured by ordinary computer program technology. In this embodiment, the case where the image evaluation calculation processing unit 13 relating to graininess is realized by computer software technology has been described. However, it is needless to say that a part or all of the image evaluation calculation processing unit 13 may have a hardware configuration using an individual circuit. It is. Next, with reference to FIGS. 5 and 6, an example of a result of actually using the above-described image graininess evaluation method showing image roughness of the present embodiment by the image evaluation value inspection unit 135 will be described. explain.

FIG. 5 is a diagram showing the result of showing the image evaluation value according to the present embodiment as a relationship with each image sample, and FIG. 6 is a diagram showing the subjective evaluation (degree of roughness) by a plurality of subjects. It is.

An image evaluation value indicating the degree of image roughness was calculated for a large number of digital image samples prepared by a silver halide photograph, an ink jet printer and a sublimation printer, and a sensory evaluation was also performed. All the image samples to be measured were white / black images, and a halftone image having a Macbeth densitometer having an image density of about 0.9 to 1.0, which is the most noisy white / black image in the density range, was used.
In addition, the subjective sensory evaluation was performed by about 10 subjects, and the roughness level of each image sample was ranked, and the subjective evaluation value was used as an average of the results. FIG. 6 shows the result.

The image evaluation value inspection means 135 compares the standard image sample representing the subjective evaluation value and the image evaluation sample 3 representing the image evaluation value with an optical magnifying means (not shown) and an illumination device. (Not shown), the standard image sample and the image evaluation sample 3 are visually inspected, and the accuracy of the image evaluation value can be simply compared and inspected.
As can be seen by comparing the results in FIGS. 5 and 6,
The image evaluation value according to the present embodiment shown in FIG.
The results according to the present embodiment are very similar to the results of the sensory evaluation shown in FIG. 2. According to the method according to the present embodiment, the tendency of the sensory evaluation can be quantitatively observed.

As described above, according to the digital image image evaluation apparatus according to the present embodiment, the image information input means,
Image information storage means, preprocessing means, Wiener spectrum calculation means, spatial frequency characteristic correction means, by including the image evaluation value inspection means, when a digital image is measured by a Macbeth densitometer or the like, It is possible to realize a digital image image evaluation apparatus capable of evaluating the image quality visually perceived in a region having a certain image density, that is, a graininess.

[0030]

As described above, according to the digital image image evaluation apparatus according to the first aspect of the present invention, when a halftone is expressed in the evaluated image, which is a digital image, the degree of roughness felt visually is considered. Combining the psychological evaluation that expresses the degree of graininess with the physical evaluation that evaluates the degree of graininess of the image with a numerical value that is objectively measured, an image of a digital image that evaluates the graininess perceived as graininess in an arbitrary halftone digital image In the evaluation device, by making the result of the physical evaluation close to the result of the psychological evaluation, a one-dimensional Wiener spectrum of an image having a certain density is obtained by a one-dimensional Fourier analysis method. This is subjected to discrete Fourier transform data processing, and an image evaluation value is calculated for the roughness of the image. Exhibit valuable and highly correlated, it is possible to conventional digital image which could not be realized by the evaluation method of the image, the evaluation of graininess certain images. Further, by considering the product of the two parameters of the spatial frequency characteristic and the standard deviation of the image density, the variation of the image density and the space (space between adjacent pixels) where the dots forming the image are arranged are considered. By introducing an analysis method that considers human visual characteristics as a total of variation, it is corrected to reflect the one-dimensional characteristics of the human visual system,
The image evaluation value becomes more reliable.

According to a second aspect of the present invention, there is provided a digital image image evaluation apparatus according to the first aspect of the present invention, wherein the digital image image evaluation apparatus includes a digital image image evaluation apparatus based on image density data captured from a halftone digital image. By performing a discrete one-dimensional Fourier transform, obtaining a Wiener spectrum of the digital image density data, and performing quantitative evaluation from two parameters of the Wiener spectrum and the variation of the density data, The same effect as that of the digital image evaluation apparatus according to the first aspect can be obtained.

According to a third aspect of the present invention, in the digital image image evaluating apparatus according to the first aspect of the present invention, the image to be evaluated includes an optical density level and a one-dimensional image. Image information input means for inputting as one-dimensional image information including dynamic position information; image information storage means for storing the one-dimensional image information input by the one-dimensional image information input means; Preprocessing means for applying a one-dimensional Fourier transform, and a winner for generating information indicating a spatial frequency distribution of one-dimensional image density data by calculating a Wiener spectrum of a Fourier transform which is an output of the preprocessing means The spectral frequency calculating means and the spatial frequency at which the sensitivity of the spatial frequency characteristic of the human visual system becomes maximum with respect to the output of the Wiener spectrum calculating means; And the value of the Wiener spectrum which,
A spatial frequency characteristic correction unit that corrects a spatial frequency by multiplying by a standard deviation of image density data, and an optical magnifying unit and an illumination unit, and a known standard image that the spatial frequency characteristic correction unit has, Compare and inspect the image to be evaluated,
By providing the image evaluation value inspection means for simply comparing and evaluating the accuracy of the image evaluation value, the same effect as that of the digital image image evaluation apparatus according to claim 1 can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a main configuration of an image evaluation device according to an embodiment.

FIG. 2 is a block diagram illustrating a configuration of an image evaluation value calculation processing unit of the image evaluation device in FIG. 1;

FIG. 3 shows visual spatial frequency characteristics.

FIG. 4 is a diagram illustrating an example of a power spectrum of image density data.

FIG. 5 is a diagram showing a result of showing an image evaluation value according to the present embodiment as a relationship with each image sample.

FIG. 6 is a diagram showing a subjective evaluation (degree of roughness) by a plurality of subjects.

FIG. 7 is a diagram illustrating a result of measuring a change in an evaluation value depending on a scanning angle when a conventional image noise evaluation method is applied to evaluation of a digital image.

[Explanation of symbols]

 1: Personal computer 2: Printing means 3: Image evaluation sample 4: Micro densitometer 5: Micro spot irradiating means 6: Spot 7: Image reading area 8: Image information input means 9: Image signal reflectance detection unit 10: Image density output Part 11: Micro density value 12: One-dimensional position information 13: Image evaluation calculation processing part 131: Image density memory (image information storage means) 132: Preprocessing means 133: Wiener spectrum calculation means 134: Spatial frequency characteristic correction Means 135: Image evaluation inspection processing unit 14: Operation processing procedure storage unit 15: Operation control unit 16: Operation result output unit

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B057 AA11 BA24 CA07 CA11 CB07 CB11 CE13 CH09 CH11 DA01 DA04 DA08 DB01 DB09 5C077 LL11 LL20 MP01 MP06 PP03 PP09 PP15 PP43 PP49 PQ20 PQ22 SS02 TT05 5L096 AA06 BA03 FA07 FA17 FA43 MA01

Claims (3)

[Claims] When a halftone is expressed in an image to be evaluated, which is a digital image, a psychological evaluation indicating a degree of roughness perceived visually, and a physics for evaluating the degree of roughness of an image with a numerical value measured objectively. In the digital image image evaluation device for evaluating the graininess perceived as roughness in an arbitrary halftone digital image by combining the evaluation with the evaluation, the result of the physical evaluation is close to the result of the psychological evaluation. An image evaluation device for digital images, characterized in that: 2. The digital image image evaluation apparatus according to claim 1, wherein a discrete one-dimensional Fourier transform is performed based on the image density data captured from the halftone digital image, and the digital image density data winner is obtained. An image evaluation apparatus for a digital image, wherein a spectrum is obtained, and quantitative evaluation is performed based on two parameters of the Wiener spectrum and the variation of density data. 3. The digital image image evaluation device according to claim 1, wherein the image to be evaluated is input as one-dimensional image information including an optical density level and one-dimensional position information. Means, image information storage means for storing the one-dimensional image information input by the one-dimensional image information input means, preprocessing means for performing a discrete one-dimensional Fourier transform on the image information, The output Fourier transform winner
Calculating a spectrum to generate information indicating a spatial frequency distribution of the one-dimensional image density data; and a Wiener / Spectrum calculating means. A spatial frequency characteristic correction means for multiplying the value of the Wiener spectrum with respect to the spatial frequency at which the sensitivity of the maximum is obtained and the standard deviation of the image density data to correct the spatial frequency, an optical magnification means and an illumination means, A known standard image of the spatial frequency characteristic correction unit and the image to be evaluated are compared and inspected, and an image evaluation value inspection unit for easily comparing and evaluating the accuracy of the image evaluation value is provided. An image evaluation device for digital images.