KR101024104B1 - METHOD AND APPARATUS OF ASSESSING IMAGE QUALITY USING THE QUANTIZATION IN THE R-θ PLANE - Google Patents

Abstract

PURPOSE: A method and an apparatus for evaluating the pixel quality by partitioning a polar coordinate plane into a quantized region are provided to efficiently extract an image in an RR pixel quality evaluation method. CONSTITUTION: An original image and a deteriorate image are partitioned to a plurality of unit blocks(410). Each pixel value of the unit blocks is calculated(420). A representative image pixel difference between a horizontal direction and a vertical direction is calculated by block(430). The representative pixel value difference is transformed to the coordinate of a polar coordinate plane(440). An image quality assessment metric is calculated by using the difference between the original image and the deteriorate image(460).

Description

Method and apparatus for assessing image quality using polar plane segmentation {Method and apparatus of assessing image quality using the quantization in the r-θ plane}

The present invention relates to image quality evaluation using polar plane splitting, and more particularly, to a quality evaluation method using polar plane splitting, which is more similar to a subjective image quality evaluation result directly evaluated by a person than a conventional objective image quality evaluation method. Relates to a device.

Image quality assessment (IQA) is a measure of how good an image is. Image quality evaluation is divided into two types: subjective quality evaluation and objective quality evaluation. Subjective image quality evaluation is more accurate than objective image quality evaluation because the quality of still images or videos is evaluated through human visual perception. However, this subjective image quality evaluation is not suitable for use in actual applications because of the problem of repeatedly making expensive evaluations. Therefore, many objective image quality evaluations have been studied because the objective image quality evaluation is more suitable for the actual image quality evaluation than the subjective image quality evaluation. In general, the closer the objective quality evaluation result is to the subjective quality evaluation result, the more accurate image quality evaluation.

Recently, many image quality evaluation methods have been developed, and VQEG (Video quality experts group) and ITU (International Telecommunication Union) play a role in setting standards. LIVE (Laboratory for image and video engineering) is studying the objective image quality evaluation method of still images and videos, and the image database provided by LIVE is used by many researchers.

The objective image quality evaluation method is based on the difference between the original image and the degraded image. There are three categories of objective image quality evaluation methods: full reference (FR), reduced reference (RR), and no-reference (NR).

In the FR, it is assumed that all the information of the image is known, and thus the most accurate results are obtained and more research is conducted than the other two objective image quality evaluation methods. Representative full reference (FR) image quality evaluation methods include mean square error (MSE) and peak signal to noise ratio (PSNR). However, these methods do not accurately reflect the human visual system (HVS).

Structural similarity (SSIM) evaluates the image quality by calculating the average, variance, and covariance between the original image and the degraded image, and uses the structural difference of the image. For example, a single value decomposition (SVD) index is a method of comparing singular values of each block in an original image and a degraded image. However, since the original image is not always available in most practical applications, in such a case, a referenceless (NR) quality evaluation method can be used. The NR image quality evaluation method does not use any original image to evaluate image quality. Instead, they use artifacts such as color bleeding, blocking, and blurring. The NR image quality evaluation method has the advantage of not needing to transmit data since no information on the original image is used, but it is not as effective as the image quality evaluation method. On the other hand, unlike the FR image quality evaluation method, the features extracted from the image are used in the RR image quality evaluation method.

1 is a block diagram of a conventional RR image quality evaluation apparatus.

Referring to FIG. 1, a feature extractor of a transmitter extracts a feature from an original image and transmits the extracted feature to a receiver through an auxiliary channel. The receiver evaluates the image quality of the deteriorated image received by the receiver through the deterioration channel unit by using the feature received through the auxiliary channel. That is, the receiver evaluates the image quality by comparing the feature extracted from the original image received through the degradation channel unit and the feature received from the transmitter through the auxiliary channel unit in the RR image quality analyzer. The feature extracted from the original image consists of a much smaller amount of data than the total amount of original image. Therefore, the performance of the RR quality estimation method depends on the features extracted from the original image. Therefore, it is necessary for the RR image quality evaluation method to effectively extract a feature suitable for image quality evaluation.

Accordingly, the first problem to be solved by the present invention is to provide an image quality estimation method using polar coordinate plane division that can effectively extract the features of the image used in the RR image quality evaluation method.

A second problem to be solved by the present invention is to provide an image quality evaluation apparatus using polar coordinate plane division that can effectively extract the features of the image used in the RR image quality evaluation method.

The present invention comprises the steps of: dividing the original image and the degraded image into a plurality of unit blocks to achieve the first object; Calculating a representative pixel value of each of the divided unit blocks; Calculating a representative pixel value difference in a horizontal direction and a vertical direction for each unit block; Converting the calculated difference between the representative pixel values in the horizontal direction and the vertical direction into coordinates on a polar coordinate plane; Calculating the number of regions including the area corresponding to each of the quantized regions in which the coordinates on the converted polar coordinate plane belong to each of the original image and the deteriorated image when the polar coordinate plane is divided into quantized regions; Calculating a difference of the area inclusion count between the original image and the deteriorated image for each quantization region; And calculating an image quality evaluation index based on the calculated difference of the area inclusion counts for each quantization region.

According to an embodiment of the present invention, the pixel value may be a brightness value.

In addition, the quantization regions of the polar planes r and θ may be formed by uneven quantization of the r-axis and uniform quantization of the θ-axis.

According to another embodiment of the present invention, when the area inclusion counts belonging to each of the quantization areas are calculated for each of the original image and the deteriorated image, a histogram of the quantization areas as a class interval and the frequency of the area inclusions as the frequency is calculated. It may be calculated for each image and the deteriorated image.

According to another embodiment of the present invention, the difference in the area inclusion count between the original image and the deteriorated image may be calculated for each quantization region, and the difference in the area inclusion count for each quantization region may be summed.

The calculating of the representative pixel value of each of the divided unit blocks may be performed by calculating the average of the brightness values of the pixels included in the divided unit blocks as the representative pixel value. The unit block is preferably an 8 × 8 unit block.

The present invention provides a block divider for dividing an original image and a degraded image into a plurality of unit blocks in order to achieve the second object. A representative pixel value calculator configured to calculate a representative pixel value of each of the divided unit blocks; A representative pixel value difference calculator for calculating a difference between the representative pixel values in the horizontal direction and the vertical direction for each unit block; A polar coordinate converter configured to convert the calculated difference between the representative pixel values in the horizontal direction and the vertical direction into coordinates on a polar coordinate plane; A quantization unit configured to calculate, according to the original image and the deteriorated image, the number of inclusions of the region in which the coordinates on the converted polar coordinate plane belong to each of the quantization regions when the polar coordinate plane is divided into quantization regions; And an image quality evaluation index calculator for calculating a difference in the area inclusion frequency between the original image and the deteriorated image for each quantization region, and calculating an image quality evaluation index by using the calculated difference in the area inclusion frequency. Provided is an image quality evaluation apparatus using polar coordinate plane division.

According to the present invention, since the feature of the image used in the RR image quality estimation method can be effectively extracted by dividing the polar plane into a quantized region, it can be used as an excellent RR image quality index in providing multimedia services. . In addition, according to the present invention, it is more similar to the subjective image quality evaluation result directly evaluated by the human person than the evaluation result by the conventional objective image quality evaluation method, showing a better image quality performance.

EMBODIMENT OF THE INVENTION Hereinafter, this invention is demonstrated in detail.

According to an embodiment of the present invention, a method for evaluating image quality using polar coordinate plane segmentation comprises: dividing an original image and a degraded image into a plurality of unit blocks; Calculating a representative pixel value of each of the divided unit blocks; Calculating a representative pixel value difference in a horizontal direction and a vertical direction for each unit block; Converting the calculated difference between the representative pixel values in the horizontal direction and the vertical direction into coordinates on a polar coordinate plane; Calculating the number of inclusions of the region in which the coordinates on the converted polar coordinate plane belong to each of the quantization regions when the polar coordinate plane is divided into quantized regions for each of the original image and the deteriorated image; Calculating a difference of the area inclusion count between the original image and the deteriorated image for each quantization region; And calculating an image quality evaluation index based on the calculated difference in region inclusion counts for each quantization region.

The image quality estimation method using polar coordinate plane division according to an embodiment of the present invention is a reduced-reference (RR) image quality evaluation method and is based on a vector quantization histogram quality metric (VQHQM).

Hereinafter, the present invention will be described in more detail with reference to preferred examples. However, these examples are intended to illustrate the present invention in more detail, it will be apparent to those skilled in the art that the scope of the present invention is not limited thereby.

2 is a block diagram of an image quality evaluation apparatus using polar coordinate plane division according to an embodiment of the present invention.

Referring to FIG. 2, the apparatus for evaluating image quality according to an embodiment of the present invention may include a block divider 210, a representative pixel value calculator 220, a representative pixel value difference calculator 230, a polar coordinate converter 240, and quantization. The unit 250 includes a picture quality evaluation index calculator 260 and a picture quality evaluation index calculator 270.

The block dividing unit 210 receives the original image and the degraded image, respectively, and divides the received original image and the degraded image into a plurality of unit blocks. Such unit blocks are preferably 8 × 8 unit blocks.

The representative pixel value calculator 220 calculates a representative pixel value of the pixels in the unit block divided by the block divider 210. In this case, the representative pixel value is preferably an average value of brightnesses of the pixels in the unit block, but a method of determining the representative pixel value may be determined using various other methods. When the original image and the degraded image are divided into 8 × 8 unit blocks, and the divided 8 × 8 unit blocks are replaced with the average brightness values of the pixels in the unit block, the amount of calculation is reduced and the dimension space of the feature space is reduced. It has a decreasing effect. In addition, although one embodiment of the present invention considers brightness as a pixel value, other values representing the pixel may be, for example, hue, saturation, and brightness.

The representative pixel value difference calculator 230 calculates the representative pixel value difference in the horizontal direction and the vertical direction for each unit block. According to an exemplary embodiment of the present invention, the representative pixel value difference calculator 230 calculates a difference between average brightness values in the horizontal direction and the vertical direction for each 8 × 8 unit block. The representative pixel value difference calculator 230 calculates a difference of average brightness values between adjacent 8 × 8 unit blocks in each of the horizontal direction and the vertical direction by using Equation 1 below.

Where dI _x is the difference between the average brightness values calculated along the horizontal direction of the image in the unit block located at (i, j), and dI _y is calculated along the vertical direction of the image in the unit block located at (i, j). The difference is the average brightness value. The (dI _x , dI _y ) pairs calculated in each unit block may be represented as one variation vector in the dI _x -dI _y coordinate plane.

The distribution of (dI _x , dI _y ), which is the result of calculating the brightness difference value along the horizontal direction and the vertical direction for the entire 8 × 8 unit blocks constituting the image, is represented by the representative pixel value difference calculator 230. The characteristics of the image will be displayed.

The polar coordinate converting unit 240 converts the coordinates (dI _x , dI _y ) calculated by the representative pixel value difference calculating unit 230 into polar coordinates (r, θ). The polar coordinate r represents the magnitude of the brightness shift between adjacent unit blocks and is expressed as a distance from the origin. Θ of the polar coordinate means the brightness shift direction. Accordingly, the polar coordinates (r, θ) converted by the polar coordinate converting unit 240 from the (dI _x , dI _y ) coordinates may be represented as a variation vector in the polar coordinates.

The quantization unit 250 quantizes the disparity vector in polar coordinates. That is, the quantization unit 250 divides the polar plane into quantization regions, and calculates the number of regions included in each of the quantization regions of the polar plane by the original vector and the degraded image. In this case, it is preferable that the quantization along the θ axis of the polar coordinates is uniformly divided, and the quantization along the r axis of the polar coordinates is uniformly divided. That is, the quantization unit 250 generates a vector quantization histogram by using quantization regions quantized in the polar coordinate plane with respect to the r and θ axes as class intervals, and the number of disparity vectors included in each of the quantized regions as the histogram frequency. can do. Vector quantization histograms are used in a variety of applications such as face recognition.

3 illustrates a method in which the quantization unit 250 quantizes the disparity vector in polar coordinates according to an embodiment of the present invention.

3 (a) shows quantized regions in the polar plane. FIG. 3 (b) shows quantization indices corresponding to the quantized regions of FIG. 3 (a) as a table.

Referring to FIG. 3 (a), the number of quantization levels for the θ axis is eight, and the quantization determination levels for the r axis are 1, 4, 7, 12, 20, and 60. Thus, the total number of regions of quantized regions on the polar coordinate plane is 42. Such a quantization level or quantization determination level is empirically determined according to an embodiment of the present invention, and may be variously determined according to a situation.

Referring to FIG. 3B, in the regions where 0 ≦ r ≦ 1 and r ≧ 60, one quantized region is formed regardless of θ (the brightness shift direction). Meaning that the quantization unit 250 quantizes the polar (r, θ) disparity vector means that each time the polar (r, θ) disparity vector is included in the divided quantized region on the polar plane, the area inclusion count corresponding to the quantization region is determined. It means to increase.

Meanwhile, the quantization unit 250 may generate histograms corresponding to the original image and the degraded image by calculating the area inclusion count of the polar coordinate (r, θ) disparity vector belonging to the quantized region. The histogram corresponding to the original image is represented by H _r , and the histogram corresponding to the degraded image is represented by H _d . Referring to FIG. 3 (b), the histogram is divided into 42 rank sections, and 84 bytes (42 × 2 bytes) of storage space are required to represent the entire image as a histogram. 84 bytes of storage is less than that of other image quality evaluation methods. 2 bytes is the number of bytes needed to represent floating point values.

The image quality evaluation index calculator 260 calculates a difference in the number of area inclusions between the original image and the deteriorated image for each quantization region, and calculates the image quality evaluation index by using the same. At this time, by using the difference between the histogram of the original image and the histogram of the degraded image generated by the quantization unit 250, it is possible to easily calculate the difference in the number of areas included. The difference in the total area inclusion count (VQH _diff ) between the histograms may be calculated as in Equation 2 below.

Where k is an index representing the quantized region and M is the total number of quantized regions. In one embodiment of the present invention, M is 42. As can be seen in Equation 2, the difference of the area inclusion count between the original image and the degraded image is calculated for each quantization region, and the results are summed to calculate the difference in the total area inclusion count (VQH _diff ) between the original image and the degraded image. do. The calculated difference in total area coverage (VQH _diff ) is a vector quantization histogram quality metric (VQHQM).

4 is a flowchart of an image quality evaluation method using polar coordinate plane division according to an embodiment of the present invention.

In operation 410, the image quality evaluation apparatus receives the original image and the degraded image, and divides the received original image and the degraded image into a plurality of unit blocks. Such unit blocks are preferably 8 × 8 unit blocks. It is preferable to remove projection noise by using a low pass filter before or after division into 8x8 blocks.

In operation 420, the image quality evaluation apparatus calculates a representative pixel value of each of the divided unit blocks in operation 410. In this case, the representative pixel value is preferably an average value of brightnesses of the pixels in the unit block, but a method of determining the representative pixel value may be determined using various other methods.

In operation 430, the image quality evaluation apparatus calculates a difference between representative pixel values in the horizontal direction and the vertical direction for each unit block. According to an embodiment of the present invention, the difference between the average brightness values in the horizontal direction and the vertical direction for each 8 × 8 unit block is calculated. In this case, the difference in the average brightness value between adjacent 8 × 8 unit blocks in each of the horizontal and vertical directions may be calculated using Equation 1 described above.

In operation 440, the image quality evaluation apparatus converts the difference in the representative pixel values in the horizontal direction and the vertical direction calculated in operation 430 from coordinates dI _x −dI _y to coordinates (r, θ) on the polar coordinate plane. The polar coordinate r indicates the magnitude of the brightness shift between adjacent unit blocks, and is expressed as a distance from the origin, and the? Of the polar coordinate indicates the brightness shift direction. Therefore, the polar coordinates (r, θ) converted from the (dI _x , dI _y ) coordinates described in Equation 1 may be expressed as a variation vector in the polar coordinates.

In operation 450, the image quality evaluation apparatus calculates, for each of the original image and the deteriorated image, the number of region inclusions in which the coordinates (r, θ) on the polar plane converted in operation 440 belong to each quantized region of the polar plane. This means disparity vector quantization in polar coordinates. When quantizing the polar plane, it is preferable that the quantization along the θ axis of the polar coordinates is uniformly divided into regions, and the quantization along the r axis of the polar coordinates is uniformly divided into the polar plane regions. In addition, a vector quantization histogram may be generated by using the quantization regions of the polar plane as the class interval, and the number of disparity vectors included in each of the quantized regions as the frequency of the histogram.

In operation 460, the image quality evaluating apparatus calculates the image quality evaluation index from the difference in the number of area inclusions between the original image and the degraded image. In this case, the difference in the number of region inclusions may be calculated by calculating the difference of each class interval between the histogram of the original image and the histogram of the degraded image generated in operation 450. The difference in the total area inclusion counts (VQH _diff ) between the histograms is calculated as in Equation 2, and the difference in the total area inclusion counts (VQH _diff ) is the vector quantization histogram quality index (VQHQM).

The image quality evaluation method using polar coordinate plane division according to an embodiment of the present invention described above is a kind of method for evaluating image quality by calculating a difference between an original image and a degraded image. There are three categories of image quality evaluation methods: full reference (FR), reduced reference (RR), and no-reference (NR). In one embodiment of the present invention, a vector quantization histogram method, which is an RR quality evaluation method, is used. ). After dividing an image into blocks, a histogram is generated by representing the amount of change in the representative pixel value of the block as a vector and summing the number of vectors belonging to the quantized region of the polar coordinate plane for each quantized region. This method of generating a histogram is called a vector quantization histogram method.

This vector quantization histogram is used to effectively evaluate the RR quality. Hereinafter, looking at the similarity between the VQHQM and the differential mean opinion score (DMOS) according to an embodiment of the present invention for LIVE (Laboratory for image and video engineering) images, and comparing with other image quality indexes Shall be.

5 shows an example of an experimental image used in an embodiment of the present invention. 5 (a) is a building image (768 × 512), FIG. 5 (b) is a parrot image (768 × 512), FIG. 5 (c) is a sailboat image (480 × 720), and FIG. 5 (d) is a lighthouse image 5 (e), FIG. 5 (e) shows a female image wearing a hat (480 x 720).

6 illustrates an example of an original image and different types of deteriorated images used in an embodiment of the present invention. Fig. 6 (a) shows the original image, Fig. 6 (b) shows the JPEG deteriorated image of the original image, Fig. 6 (c) shows the JPEG 2000 deteriorated image of the original image, and Fig. 6 (d) shows the Gaussian blur of the original image. Gaussian blur deteriorated image, FIG. 6 (e) shows white noise deteriorated image of the original image, and FIG. 6 (f) shows fast fading deteriorated image of the original image.

The purpose of evaluating image quality is to obtain quantitative data similar to subjective image quality evaluation of human beings. Therefore, in order to measure the performance of the image quality estimation method according to an embodiment of the present invention, first, mean opinion score (MOS) data of an experimental image is required. Mean opinion score (MOS) data reflects the human visual system (HVS) because it is derived from subjective tests of human image quality. Therefore, MOS (mean opinion score) data can be used as a criterion for evaluating image quality and are necessary for comparing the performance of image quality evaluation. Differential mean opinion score (DMOS) means subjective picture quality, and DMOS can be obtained by applying mean opinion score (MOS), which is a value that a person directly evaluates picture quality, to the double-stimulus continuous quality-scale (DSCQS) method. .

In order to evaluate the performance of different image quality indexes, Pearson Correlation Coefficient (Pearson Correlation Coefficient) is used to measure the similarity between the quality index and the DMOS.

Table 1 compares Pearson's correlation coefficients of four different image quality indexes calculated for each deteriorated image. The four image quality indexes are mean squared error (MSE), structural similarity index measurement (SSIM), mean singular value decomposition (MSVD), and VQHQM, and each image quality index is calculated for five different degradation images.

The underlined values in Table 1 are the highest Pearson correlation coefficients for each of the five different deteriorated images, and the higher the Pearson correlation coefficient, the more similar it is to DMOS, which is a measure evaluated by human vision. Therefore, it can be seen that the VQHQM according to an embodiment of the present invention shows the best performance. According to an embodiment of the present invention, VQHQM shows the best performance for JPEG 2000, JPEG, white noise, Gaussian blur, and SSIM for fast fading. It has the best performance.

In addition, a total of 982 LIVE (Laboratory for Image & Video Engineering) experimental images are used to compare the performance of the image quality evaluation method and the conventional image quality evaluation method according to an embodiment of the present invention.

Table 2 is a table comparing performance by calculating Pearson's correlation coefficient for each of 982 LIVE test images by each image quality evaluation index.

Referring to Table 2, it can be seen that the image quality evaluation index indicating the largest Pearson correlation coefficient is VQHQM according to an embodiment of the present invention.

7 shows a scatter plot between the DMOS and four different picture quality indexes. Referring to FIG. 7, it can be seen that VQHQM according to an embodiment of the present invention shows better fitting results than other image quality evaluation indices.

The VQHQM generated according to an embodiment of the present invention is one of the indexes for evaluating the image quality of the degraded image. VQHQM is calculated using a histogram obtained by calculating the number of disparity vectors belonging to the quantized region. As a result of using each of the image quality index and the PMOS Pearson correlation coefficient in the performance comparison, it can be seen that the VQHQM generated according to an embodiment of the present invention shows better performance than the conventional image quality index.

Embodiments of the present invention may be implemented in the form of program instructions that can be executed by various computer means and recorded in a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on the media may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs, DVDs, and magnetic disks, such as floppy disks. Magneto-optical media, and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware device described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

As described above, the present invention has been described by specific embodiments such as specific components and the like. For those skilled in the art to which the present invention pertains, various modifications and variations are possible. Therefore, the spirit of the present invention should not be limited to the described embodiments, and all of the equivalents or equivalents of the claims as well as the claims to be described later will belong to the scope of the present invention. .

So far I looked at the center of the preferred embodiment for the present invention. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

1 is a block diagram of a conventional RR image quality evaluation apparatus.

2 is a block diagram of an image quality evaluation apparatus using polar coordinate plane division according to an embodiment of the present invention.

3 illustrates a method in which the quantization unit 250 quantizes the disparity vector in polar coordinates according to an embodiment of the present invention.

4 is a flowchart of an image quality evaluation method using polar coordinate plane division according to an embodiment of the present invention.

5 shows an example of an experimental image used in an embodiment of the present invention.

6 illustrates an example of an original image and different types of deteriorated images used in an embodiment of the present invention.

7 shows a scatter plot between the DMOS and four different picture quality indexes.

Claims (15)

Dividing the original image and the degraded image into a plurality of unit blocks; Calculating a representative pixel value of each of the divided unit blocks; Calculating a representative pixel value difference in a horizontal direction and a vertical direction for each unit block; Converting the calculated difference between the representative pixel values in the horizontal direction and the vertical direction into coordinates on a polar coordinate plane; Calculating the number of regions including the area corresponding to each of the quantized regions in which the coordinates on the converted polar coordinate plane belong to each of the original image and the deteriorated image when the polar coordinate plane is divided into quantized regions; Calculating a difference of the area inclusion count between the original image and the deteriorated image for each quantization region; And And calculating a picture quality evaluation index based on the calculated difference in area inclusion counts for each quantization region. The method of claim 1, And the pixel value is a brightness value. The method of claim 1, The quantization regions of the polar planes (r, θ) are formed in such a way that the r-axis is non-uniformly quantized and the θ-axis is uniformly quantized. The method of claim 1, Computing the area inclusion counts belonging to each of the quantization areas for each of the original image and the deterioration image And calculating a histogram having the frequency of the region included in the quantized regions as the class interval for each of the original image and the deteriorated image. The method of claim 1, The step of calculating the difference of the area inclusion count between the original image and the deterioration image for each quantization region is And summing the difference of the area inclusion counts for each quantization region between the original image and the deteriorated image. The method of claim 1, Computing the representative pixel value of each of the divided unit blocks And calculating an average of brightness values of pixels included in the divided unit blocks as a representative pixel value. The method of claim 1, And the unit block is an 8 × 8 unit block. A computer-readable recording medium having recorded thereon a program for executing the method of claim 1 on a computer. A block dividing unit dividing the original image and the degraded image into a plurality of unit blocks; A representative pixel value calculator configured to calculate a representative pixel value of each of the divided unit blocks; A representative pixel value difference calculator for calculating a difference between the representative pixel values in the horizontal direction and the vertical direction for each unit block; A polar coordinate converter configured to convert the calculated difference between the representative pixel values in the horizontal direction and the vertical direction into coordinates on a polar coordinate plane; A quantization unit configured to calculate, according to the original image and the deteriorated image, the number of inclusions of the region in which the coordinates on the converted polar coordinate plane belong to each of the quantization regions when the polar coordinate plane is divided into quantization regions; And And a picture quality evaluation index calculator configured to calculate a difference in the area inclusion frequency between the original image and the deteriorated image for each quantization area, and calculate a picture quality evaluation index using the calculated difference in the area inclusion frequency. Image quality evaluation device using plane segmentation. The method of claim 9, And the pixel value is a brightness value. The method of claim 9, The quantization regions of the polar plane are quantized nonuniformly in the r-axis, and the θ axis is uniformly quantized. The method of claim 9, The quantization unit And a histogram having the frequency of the region included in the quantized regions as the class interval for each of the original image and the deteriorated image. The method of claim 9, The image quality evaluation index calculation unit And an image quality evaluation index is calculated by summing differences in the area inclusion counts between the original image and the deteriorated image for each quantization region. The method of claim 9, The representative pixel value calculating unit And an average of brightness values of pixels included in the divided unit blocks as a representative pixel value. The method of claim 9, And the unit block is an 8 × 8 unit block.

Images