JP2006174182A - Image estimation device, image estimation method, and its recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image estimation device, an image estimation method, and a recording medium which can easily and precisely quantify and estimate quality (picture quality) especially a variation which is one of abnormal images. <P>SOLUTION: An image input part 3 reads a pattern for the estimation of an image to be estimated outputted from an image forming device 20 as image data, and an Euclidean distance calculation part 6 calculate an Euclidean distance. Next, a maximum value calculation part 7 calculates a maximum value excluding noise pixel components which are scattered in a peripheral part of an image feature amount from the all Euclidean distances obtained from the Euclidean distance calculation part 6. The maximum value of the calculated image feature amount of the image to be estimated is estimated from a linear relation with a subjective estimation result (a subjective estimation point) preliminarily stored in an estimation part 8, and the estimation result of the image quality is outputted from an estimation result output part. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an image evaluation apparatus for evaluating the quality (image quality) of an image formed by an image forming apparatus such as a copying machine or a printer, an image evaluation method, and a recording medium on which a program applicable to the image evaluation apparatus is recorded.

In the process of forming an image by the electrophotographic method, the omission of an image that is caused by a transfer defect or the like and the portion to be originally transferred is not completely transferred is referred to as “bottom”. This is a kind of abnormal image in image formation and is a cause of image degradation.
The quality of the image by “Bosotsuki” is judged by visual inspection, and it is a sensual evaluation that prepares a limit sample and class sample in the visual inspection and ranks based on the rank sample. Yes.

  In this way, in the visual evaluation method performed by preparing a limit sample or class sample, there is no quantitative evaluation standard, and evaluation results often vary depending on individual differences, environment, physical condition, etc. for sensory evaluation, It is a problem. Further, since the evaluation is performed for a long time, the eyes are overworked, and there is a problem that the burden on the inspector is very large.

As an invention for solving the above problems, an image quality evaluation apparatus and an image quality evaluation method capable of accurately evaluating graininess with simple processing (for example, see Patent Document 1), and image quality of a final printed matter There are those that provide a printing apparatus including an image reading system and a printing system that can be easily adjusted (see, for example, Patent Document 2).
Japanese Patent Laid-Open No. 11-164081 JP-A-9-009088

  However, the above documents have the following problems.

  The blur evaluation method described in Patent Document 1 reads a printed product of a horizontal line image via an image reading system. This is a method of binarizing the read image and counting the number of broken horizontal lines in the binary image to obtain a blur amount. With this method, it is impossible to evaluate the quality of the blur appearing in the solid image pattern unique to the color image. In addition, there is a problem that the color image cannot be quantified with high accuracy.

  In addition, since the image quality evaluation method described in Patent Document 2 is based on binarization processing for dot images, it is still impossible to evaluate a single color solid image or a secondary color solid image when colorants are superimposed. There was a problem.

  The present invention has been made in view of the above problems, and does not require a large-scale system or complicated processing, and the image quality for a solid image of an arbitrary color including a secondary color formed by an image forming apparatus ( It is an object of the present invention to provide an image evaluation apparatus, an image evaluation method, and a recording medium capable of quantifying and evaluating a blur, which is one of abnormal images, easily and accurately.

  According to a first aspect of the present invention, there is provided an image input means for reading an image to be evaluated, which is an evaluation object formed in the image forming apparatus, as image data, and a memory for storing an evaluation criterion for determining an evaluation for the image to be evaluated. Means, a calculation means for calculating a Euclidean distance between coordinates corresponding to the pixel data based on at least two pieces of pixel data constituting the image data, a predetermined threshold, and the calculation means Features of the image to be evaluated based on the selection means for comparing the Euclidean distance and selecting the smaller one, the evaluation criterion, and one of the threshold value and the Euclidean distance selected by the selection means Detecting means for detecting an amount, and evaluating an image quality of the image to be evaluated based on the feature amount.

  A second aspect of the present invention is the image evaluation apparatus according to the first aspect, wherein the threshold value can be set by a user.

  A third aspect of the present invention is the image evaluation apparatus according to the first or second aspect, further comprising color conversion means for converting a predetermined color space in the image data into an arbitrary color space. .

  According to a fourth aspect of the present invention, in the image evaluation apparatus according to the third aspect, the color conversion unit is an RGB conversion unit that converts a predetermined color space in the image data into an RGB color space. And

  According to a fifth aspect of the present invention, in the image evaluation apparatus according to the third aspect, the color conversion means is an XYZ conversion means for converting a predetermined color space in the image data into an XYZ color space. And

  According to a sixth aspect of the present invention, in the image evaluation apparatus according to the third aspect, the color conversion unit converts the predetermined color space in the image data into a CIEL * A * B * color space. * B * conversion means.

  According to a seventh aspect of the present invention, an image to be evaluated, which is an evaluation target formed in the image forming apparatus, is read as image data, and the pixel data is handled based on at least two pieces of pixel data constituting the image data. The Euclidean distance between the coordinates is calculated, the predetermined threshold value is compared with the Euclidean distance calculated by the calculating means, the smaller one is selected, and the evaluation criterion value for the evaluation image stored in advance is selected. And a feature amount of the image to be evaluated based on the threshold value and one selected from the Euclidean distance, and an image quality of the image to be evaluated is evaluated based on the feature amount. To do.

  The invention according to claim 8 is the image evaluation method according to claim 7, wherein the threshold value can be set by a user.

  The invention according to claim 9 is the image evaluation method according to claim 7 or 8, further comprising a color conversion step of converting a predetermined color space in the image data into an arbitrary color space. .

  The invention described in claim 10 is the image evaluation method according to claim 9, wherein the color conversion step is an RGB conversion step of converting a predetermined color space in the image data into an RGB color space. And

  The invention described in claim 11 is the image evaluation method according to claim 9, wherein the color conversion step is an XYZ conversion step of converting a predetermined color space in the image data into an XYZ color space. And

  According to a twelfth aspect of the present invention, in the image evaluation apparatus according to the ninth aspect, the color conversion step converts the predetermined color space in the image data into a CIEL * A * B * color space. * B * conversion process.

  According to a thirteenth aspect of the present invention, an image to be evaluated, which is an evaluation target formed in the image forming apparatus, is read as image data, and the pixel data is handled based on at least two pieces of pixel data constituting the image data. The Euclidean distance between the coordinates is calculated, the predetermined threshold value is compared with the Euclidean distance calculated by the calculating means, the smaller one is selected, and the evaluation criterion value for the evaluation image stored in advance is selected And evaluating the image quality of the evaluated image based on the feature amount based on the threshold value and the selected one of the Euclidean distance and detecting the feature amount of the evaluated image It is a computer-readable recording medium that records a program to be executed by the apparatus.

  According to the image evaluation apparatus, the image evaluation method, and the recording medium of the present invention, it is possible to easily and accurately evaluate the quality (image quality) of an image by quantifying and evaluating one of abnormal images. It becomes.

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

FIG. 1 is a diagram showing a flow of image evaluation in the present embodiment.
An image to be evaluated is output from the image forming apparatus 20 as an evaluation image evaluation pattern. The image evaluation apparatus 1 evaluates the image quality of the output / formed image to be evaluated (hard copy, printed matter).

FIG. 2 is a block diagram showing the configuration of the image evaluation apparatus of this embodiment.
The image evaluation apparatus 1 in this embodiment includes a CPU 2, an image input unit 3, a memory 4, a color conversion unit 5, a Euclidean distance calculation unit 6, a maximum value calculation unit 7, an evaluation unit 8, and an image feature. An amount calculation unit 9, an evaluation result output unit 10, and an image bus 11 are provided.

  The CPU 2 controls the operation of the entire apparatus. The image input unit 3 reads an evaluation image such as a CCD camera, a scanner, or a scanning micro densitometer as an RGB digital signal. The memory 4 is used as a work memory for the CPU 2 and stores image data input from the image input unit 3 and measurement results. The color conversion unit 5 converts RGB image data of each pixel into a physical quantity such as a CIELAB color system (L, a, b) color value recommended by the International Illumination Commission (CIE) as a color space. The image feature amount calculation unit 9 calculates the image feature amount of the evaluation target image to be evaluated. The evaluation result output unit 10 externally outputs the evaluation result to a monitor, a printer, a file, or the like.

The image feature amount calculation unit 9 further includes a Euclidean distance calculation unit 6, a maximum value calculation unit 7, and an evaluation unit 8.
The Euclidean distance calculation unit 6 calculates the Euclidean distance starting from all the pixels of the RGB image data or some of the pixels. The maximum value calculation unit 7 calculates a maximum value from all the Euclidean distances obtained from the Euclidean distance calculation unit 6 by removing noise pixel components scattered around the image feature amount by the noise removing unit. The evaluation unit 8 evaluates the image quality of the evaluation image output from the image forming apparatus 20 to be evaluated based on the maximum value of the image feature amount extracted from the maximum value calculation unit 7.

  Next, the operation of the image evaluation apparatus of this embodiment will be described with reference to the drawings. FIG. 3 is a flowchart showing the operation of the image evaluation apparatus of this embodiment. In this embodiment, the image data of the image to be evaluated is used as RGB image data using R (red), G (green), and B (blue) that are the three primary colors of light.

  First, the evaluation pattern of the image to be evaluated output from the image forming apparatus 20 is read as RGB image data by the image input unit 3 (step S101). Next, the image feature amount calculation unit 9 calculates the image feature amount of the read RGB image data. Actually, first, the Euclidean distance calculation unit 6 calculates the Euclidean distance starting from all the pixels of the RGB image data or some of the pixels (step S102).

  Here, the maximum value calculation unit 7 calculates the maximum value from all the Euclidean distances obtained from the Euclidean distance calculation unit 6 excluding pixel components (noise components) scattered around the image feature amount. (Step S103). This is because, if the Euclidean distance of pixel components scattered around the image feature amount is included as the image feature amount of the image to be evaluated, only the pixel components that are scattered are image features to be evaluated. It will be calculated as a quantity. In other words, the portion that is not actually a blurred image feature amount is included as a so-called error component and is calculated as an image feature amount.

As a method for removing the pixel components (noise components) scattered around the image feature amount in the maximum value calculation unit 7, first, an image of the evaluation image calculated by the Euclidean distance calculation unit 6 is used. The Euclidean distance in the feature amount is required. Incidentally, as a generally known formula for calculating the Euclidean distance, for a non-negative integer n, two points a = (a ₁ , a ₂ ,..., _An ), The distance between b = (b ₁ , b ₂ ,..., b _n ) is expressed using the following formula (1).

  Here, as a method of removing the noise component described above, the Euclidean distance in the image feature amount of the image to be evaluated is used. Specifically, it is expressed using the following mathematical formula (2).

  The Euclidean distance calculated using Expression (2) is expressed as a noise component scattered around the image feature amount. That is, the value of the Euclidean distance calculated by Expression (2) is removed. Note that the value of x in Equation (2) is a non-negative integer and can be set by the user.

In this way, the maximum value of the image feature amount of the image to be evaluated calculated by the maximum value calculation unit 7 is evaluated by the evaluation unit 8 (step S104). The evaluation unit 8 stores in advance a linear relationship between the maximum value of the image feature amount of the image to be evaluated (the blurred image) calculated by the maximum value calculation unit 7 and the subjective evaluation result (subjective evaluation point). This linear relationship is a linear relationship as shown in FIG.
When the maximum value of the image feature amount is calculated from the maximum value calculation unit 7, the evaluation unit 8 calculates the evaluation target image from the linear relationship between the maximum value of the image feature amount of the blurred image and the subjective evaluation result (subjective evaluation point). Determine the subjective score. Finally, the image quality evaluation result is output from the evaluation result output unit (step S105).

  Further, an image to be evaluated (evaluation pattern) for measuring the amount of deterioration of the image due to the blur is, for example, the entire surface for each color of black (B), cyan (C), magenta (M), and yellow (Y). A solid image pattern, a checker pattern in which a solid area (solid image area) and a ground area (background image area) are alternately present, or a solid area patch portion (solid image area) at an arbitrary position on the paper ( A pattern having a plurality of points on the upper, lower, left, right, and center portions may be used.

FIG. 4 is a diagram illustrating an example of a case where the image input unit 3 reads the entire solid output image when the evaluation target image output from the image forming apparatus 20 is blurred.
FIG. 4 shows a case where the color material forming the image is cyan. For example, if the gradation value of the image input unit 3 is 8 bits, the values of R (red), G (green), and B (blue) that are the three primary colors of light are (0) as representative color values. , 255, 255).

  FIG. 5 is a graph in which an image to be evaluated output from the image forming apparatus 20 is read by the image input unit 3 and the image data is graphed. Here, the image data of the image to be evaluated is converted into an RGB image using R (red), G (green), and B (blue), which are the three primary colors of light, and the image data is converted into 3 RGB colors. It is shown as a dimensional graph.

  In the RGB color space, the Euclidean distance (d1... Dn (n is the total number of pixels)) starting from all pixels or a part of each pixel is calculated by the Euclidean distance calculation unit 6. In the calculated Euclidean distance, the maximum value of all the inter-pixel distances is detected as a blur amount.

  6 and 7 are diagrams showing the three-dimensional graph in the RGB color space shown in FIG. 5 in an easy-to-understand manner according to the level of the evaluation image. FIG. 6 is a graph showing a case where the level of blur of the image to be evaluated is high (bad). FIG. 7 is a diagram illustrating a case where the level of blur of the image to be evaluated is low (good).

  As shown in FIG. 6, when the level of blur of the image to be evaluated is poor, the pixel distribution in the RGB color space (portion surrounded by an ellipse in FIG. 6) is widened. On the other hand, as shown in FIG. 7, when the level of blur of the image to be evaluated is good, the pixel distribution in the RGB color space is narrow and dense.

FIG. 8 is a histogram showing an example of the Euclidean distance obtained by the Euclidean distance calculation unit 6. The linear relationship represented by this histogram is stored in the evaluation unit 8 in advance.
In FIG. 8, the image feature amount in the pixel distribution in the RGB color space is shown at the distance on the horizontal axis. Further, at the top (right end) of the distance on the horizontal axis, there is a Euclidean distance of pixel components scattered around the image feature amount. If the maximum value calculation unit 7 calculates up to the interspersed pixel components as the maximum value of the image feature amount, an error component will be included.

  Therefore, the maximum value calculation unit 7 removes pixel components (noise components) scattered around the image feature amount described above. This time, the value of x in Equation (2) described above was set to 3 to remove noise. In FIG. 8, the range indicated by the ellipse is the noise component. Actually, it is more preferable to remove from the minimum frequency and minimum distance to the minimum frequency and maximum distance among the pixel distances between the two points calculated as the Euclidean distance.

  In the present embodiment, the color unit of the image data input from the image input unit 3 is defined as RGB, and the Euclidean distance in the color space defined by the RGB digital signal is described as the image feature amount. However, the present embodiment is not limited to this, and generally uses (X, Y, Z) values called tristimulus values, or CIE 1976 L * a, which is a kind of uniform color space that is closer to a human interval. The present invention can also be applied to a color space converted into an arbitrary color system such as * b *.

It is a figure which shows the flow of the image evaluation in this embodiment. It is a block diagram which shows the structure of the image evaluation apparatus in this embodiment. It is a flowchart which shows operation | movement of the image evaluation apparatus in this embodiment. It is a figure which shows an example at the time of reading the whole surface solid output image when the blurring generate | occur | produces in the to-be-evaluated image. It is the figure which plotted the image data of the to-be-evaluated image on RGB space. 6 is a graph showing a case where the level of blur of the image to be evaluated is high (bad) on the RGB space shown in FIG. 5. It is a figure which shows the case where the level of the blur of the to-be-evaluated image is low (good) on the RGB space shown by FIG. 6 is a histogram showing an example of the Euclidean distance obtained by the Euclidean distance calculation unit 6.

Explanation of symbols

1 Image evaluation device 2 CPU
DESCRIPTION OF SYMBOLS 3 Image input part 4 Memory 5 Color conversion part 6 Euclidean distance calculation part 7 Maximum value calculation part 8 Evaluation part 9 Image feature-value calculation part 10 Evaluation result output part 11 Image bus 20 Image form formation apparatus

Claims (13)

An image input means for reading an image to be evaluated, which is an evaluation object formed in the image forming apparatus, as image data;
Storage means for storing evaluation criteria for determining an evaluation for the image to be evaluated;
Calculation means for calculating Euclidean distance between coordinates corresponding to the pixel data based on at least two pieces of pixel data constituting the image data;
A selection unit that compares a predetermined threshold value with the Euclidean distance calculated by the calculation unit and selects a smaller one;
Detection means for detecting a feature quantity of the image to be evaluated based on the evaluation criterion and one of the threshold value and the Euclidean distance selected by the selection means;
An image evaluation apparatus characterized by evaluating an image quality of the image to be evaluated based on the feature amount.   The image evaluation apparatus according to claim 1, wherein the threshold value can be set by a user.   The image evaluation apparatus according to claim 1, further comprising color conversion means for converting a predetermined color space in the image data into an arbitrary color space.   The image evaluation apparatus according to claim 3, wherein the color conversion unit is an RGB conversion unit that converts a predetermined color space in the image data into an RGB color space.   4. The image evaluation apparatus according to claim 3, wherein the color conversion unit is an XYZ conversion unit that converts a predetermined color space in the image data into an XYZ color space.   4. The image evaluation according to claim 3, wherein the color conversion means is a CIEL * A * B * conversion means for converting a predetermined color space in the image data into a color space of CIEL * A * B *. apparatus. The image to be evaluated, which is an evaluation target formed in the image forming apparatus, is read as image data,
Based on at least two pieces of pixel data constituting the image data, a Euclidean distance between coordinates corresponding to the pixel data is calculated,
Compare the predetermined threshold value and the Euclidean distance calculated by the calculating means, and select the smaller one,
Based on the value of the evaluation criterion for the evaluation image stored in advance and the selected one of the threshold value and the Euclidean distance, the feature amount of the evaluation image is detected,
An image evaluation method characterized by evaluating an image quality of the evaluated image based on the feature amount.   The image evaluation method according to claim 7, wherein the threshold value can be set by a user.   The image evaluation method according to claim 7, further comprising a color conversion step of converting a predetermined color space in the image data into an arbitrary color space.   The image evaluation method according to claim 9, wherein the color conversion step is an RGB conversion step of converting a predetermined color space in the image data into an RGB color space.   The image evaluation method according to claim 9, wherein the color conversion step is an XYZ conversion step of converting a predetermined color space in the image data into an XYZ color space.   The image evaluation according to claim 9, wherein the color conversion step is a CIEL * A * B * conversion step of converting a predetermined color space in the image data into a color space of CIEL * A * B *. apparatus. The image to be evaluated, which is an evaluation target formed in the image forming apparatus, is read as image data,
Based on at least two pieces of pixel data constituting the image data, a Euclidean distance between coordinates corresponding to the pixel data is calculated,
Compare the predetermined threshold value and the Euclidean distance calculated by the calculating means, and select the smaller one,
Based on the value of the evaluation criterion for the evaluation image stored in advance and the selected one of the threshold value and the Euclidean distance, the feature amount of the evaluation image is detected,
A computer-readable recording medium having recorded thereon a program for causing an image evaluation apparatus to evaluate the image quality of the image to be evaluated based on the feature amount.

Images