JP5837844B2 - Pseudo contour region detection device and program - Google Patents

Description

  The present invention relates to a pseudo contour region detection device that detects, as a pseudo contour region, an image that may generate a pseudo contour when the gradation of an original image is reduced, and a program that functions as a pseudo contour region detection device. .

  2. Description of the Related Art Conventionally, a technique for performing high-quality gradation conversion by classifying an image into a contour region and a non-contour region and selectively using a smoothing filter in the contour region and the non-contour region (for example, , See Patent Document 1). Specifically, the gradation conversion processing device described in Patent Document 1 multiplies a pixel of interest and its peripheral pixels in an N-value gradation image by multiplying them with a predetermined weighting coefficient, and adds the N-value gradation image to the pixel. By making the first smoothing filter to be converted into an M-value gradation image (N <M) and the scanning aperture size or the weighting factor to be added different from each other, the smoothness is made larger than that of the first smoothing filter, and the N-value gradation is obtained. And a second smoothing filter for converting the image into an M-value gradation image. Then, the non-contour region and the contour region in the N-value gradation image are identified, and the output of the first smoothing filter is selected in the contour region. Further, the mixture ratio is calculated from the edge amount in the N-value gradation image, and a filter in which the first smoothing filter and the second smoothing filter are mixed based on the mixture ratio is generated. Select.

Japanese Patent No. 3650766

  In the technique described in Patent Document 1, smoothing can be performed in consideration of generation of a pseudo contour by detecting a gradation region as a non-contour region from an image and performing a filtering process using a smoothing filter. Further, when displaying with a reduced number of gradations on a monitor or the like, generally, a gradation area in which gradation changes are almost flat is detected, and pseudo-contour generation prevention processing is performed on this area. . Conventionally, the gradation area has been detected by a method such as simply examining a luminance gradient with neighboring pixels. However, such a method is insufficient to detect the pseudo contour with high accuracy.

  An object of the present invention made in view of such circumstances is to provide a pseudo contour region detection device and a program capable of detecting a pseudo contour region with high accuracy.

  In order to solve the above problems, a pseudo contour region detection device according to the present invention is a pseudo contour region detection device that detects, as a pseudo contour region, an image that may generate a pseudo contour when the gradation of an original image is reduced. An original image spatial frequency conversion unit that converts an original image into a spatial frequency domain and generates a spatial frequency spectrum of the original image, and a threshold value of a spatial low frequency power ratio based on a specified number of gradation reduction bits And a threshold frequency determination unit for determining a spectral power threshold value, and acquiring a spatial frequency spectrum of the original image for each phase position, and a region where a ratio of spatial low frequency power at the same phase position exceeds the spectral power threshold value of the original image An original image pseudo contour region detection unit that detects the pseudo contour region, and the threshold value determination unit decreases as the number of gradation reduction bits increases. And determining a serial spectral power threshold.

  Furthermore, in the pseudo contour region detection device according to the present invention, the spatial frequency conversion unit performs multi-resolution analysis on the original image without decimation and converts the original image into the spatial frequency region.

  Further, in the pseudo contour area detection device according to the present invention, a gradation reduction unit that generates a gradation reduced image of an original image according to the number of gradation reduction bits, and converts the gradation reduced image into a spatial frequency domain. A gradation-reduced image spatial frequency conversion unit that generates a spatial frequency spectrum of the gradation-reduced image, and obtains the spatial frequency spectrum of the gradation-reduced image for each phase position, and the ratio of the spatial low-frequency power at the same phase position A gradation-reduced image pseudo-contour region detection unit that detects a region that exceeds the spectrum power threshold as a pseudo-contour region of a tone-reduced image, a pseudo-contour region of the original image, and a pseudo-contour region of the tone-reduced image For a common area, the spatial frequency spectrum of the original image and the spatial frequency spectrum of the gradation-reduced image are compared, and the spatial frequency spectrum of the gradation-reduced image is compared. There is further comprising a, a total false contour area detection unit for detecting an area larger than a predetermined threshold value than the spatial frequency spectrum of the original image as a final pseudo outline area.

  Furthermore, in the pseudo contour region detection device according to the present invention, the gradation reduced image spatial frequency conversion unit performs multi-resolution analysis on the gradation reduced image without decimation and converts the image into the spatial frequency region. .

  In order to solve the above problems, a program according to the present invention causes a computer to function as the pseudo contour region detection device.

  According to the present invention, it becomes possible to detect a pseudo contour region generated when the gradation of an original image is reduced with high accuracy.

It is a block diagram which shows the structural example of the pseudo | simulation outline area | region detection apparatus which concerns on Example 1 of this invention. It is a figure explaining the process of the pseudo contour area | region detection part in the pseudo contour area | region detection apparatus which concerns on Example 1 of this invention. It is a block diagram which shows the structural example of the pseudo | simulation outline area | region detection apparatus which concerns on Example 2 of this invention. It is a block diagram which shows the structural example of the synthetic | combination false outline area | region detection part in the false outline area detection apparatus which concerns on Example 2 of this invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  First, the pseudo contour region detection apparatus according to the first embodiment will be described. FIG. 1 is a block diagram illustrating a configuration example of the pseudo contour region detection apparatus according to the first embodiment. As shown in FIG. 1, the pseudo contour region detection apparatus 1 includes an original image spatial frequency conversion unit 10, a threshold determination unit 20, and an original image pseudo contour region detection unit 30.

  The original image spatial frequency transform unit 10 transforms an original image into a spatial frequency domain by a method having phase information such as discrete Fourier transform using a window function, discrete cosine transform using block division, multiresolution analysis, etc. Generate a frequency spectrum. Then, the generated spatial frequency spectrum of the original image is output to the original image pseudo contour region detection unit 30.

  In order to detect a pseudo contour region with high resolution accuracy of the original image, the original image spatial frequency conversion unit 10 performs multi-resolution analysis (for example, without decimation of the original image, that is, without reducing the image size). , N-th order wavelet decomposition). If the decomposition rank n is increased, the spatial frequency band decomposition can be performed to a band sufficiently close to the direct current component. For example, when n = 4, spatial frequency band decomposition can be performed up to 1/16 of the sampling frequency in the horizontal and vertical directions of the original image.

  The threshold determination unit 20 determines a spectral power threshold that is a threshold of the ratio of spatial low frequency power based on the number of gradation reduction bits specified by the user. Then, the determined spectral power threshold value is output to the original image pseudo contour region detection unit 30. As the number of gradation reduction bits is larger, the area where the gradation changes naturally becomes an area where the gradation changes stepwise, and the possibility that a pseudo contour is generated increases. And in the area | region where a gradation changes stepwise, the ratio of spatial high frequency power increases. Therefore, the threshold value determination unit 20 determines the spectrum power threshold value to be smaller as the number of gradation reduction bits is larger in order to increase the detection accuracy of the pseudo contour region.

  The threshold determination unit 20 may store in advance a table in which the number of gradation reduction bits and the spectrum power threshold are associated with each other, and determine the spectrum power threshold based on the correspondence table. Table 1 shows an example of a correspondence table between the number of gradation reduction bits and the spectrum power threshold. The spectrum power threshold is 1 at maximum. For example, when the threshold value determining unit 20 has the correspondence table shown in Table 1 and the number of gradation reduction bits is 2 bits, the threshold value determining unit 20 determines the spectral power threshold value as 0.92.

  The original image pseudo contour region detection unit 30 acquires the spatial frequency spectrum of each frequency band generated by the original image spatial frequency conversion unit 10 for each phase position, and the ratio of the spatial low frequency power at the same phase position is the threshold value determination unit. A region exceeding the spectrum power threshold determined by 20 is detected as a pseudo contour region.

  FIG. 2 is a diagram for explaining processing of the original image pseudo contour region detection unit 30. Here, a case where the original image is subjected to the fourth-order wavelet decomposition in the original image spatial frequency conversion unit 10 without decimation and satisfying the Parseval equation to normalize the spectrum power is shown. Image A shown in FIG. 2A is an image of spatial high-frequency components (LH, HL, HH components) in the first layer. Image B is an image of the spatial high-frequency component in the second layer. Image C is an image of the spatial high-frequency component in the third hierarchy. Image D is an image of the spatial high-frequency component in the fourth layer. The image E is an image of a spatial low frequency component (LL component) in the fourth layer. The image blocks a to e are image blocks existing at the same phase position in the images A to E. When the spatial frequency conversion unit 10 performs wavelet decomposition on the original image without decimation, the size of the image blocks a to e is changed. One pixel can be used.

  Spectral regions A to O shown in FIG. 2B represent power spectrum regions of the image blocks a to e, respectively. The original image pseudo contour region detection unit 30 analyzes the power spectrum for the image blocks a to e existing at the same phase position in the images A to E, and calculates the ratio of the spatial low frequency power. The ratio of the spatial low-frequency power is the ratio of the spectrum in the spatial low-frequency band (spectrum area A) to the spectrum in the entire frequency band (spectral area A to E).

  Then, the original image pseudo contour region detection unit 30 detects, as a pseudo contour region of the original image, a region where the calculated spatial low frequency power ratio exceeds the spectrum power threshold. The pseudo contour region is, for example, a binary image in which the phase position where the spatial low frequency power ratio exceeds the spectral power threshold is 1, and the phase position where the spatial low frequency power ratio is less than the spectral power threshold is 0. Can do.

  Note that when the spatial frequency transform unit 10 performs wavelet decomposition on the original image with decimation, the image size becomes 1/2 in the vertical and horizontal directions as the decomposition rank n increases. When the image size of e is 1 pixel, the image size of the third layer image block c is 2 × 2 pixels, the image size of the second layer image block b is 4 × 4 pixels, and the first layer image The image size of the block a is 8 × 8 pixels. That is, when the original image spatial frequency conversion unit 10 performs multiresolution analysis on the original image with decimation, the original image pseudo contour region detection unit 30 determines the size of the image block existing at the same phase position in the images of each layer. It is necessary to set the value according to the wavelet decomposition rank.

  As described above, in the pseudo contour region detection device 1 according to the first embodiment, the original image spatial frequency conversion unit 10 converts the original image into the spatial frequency region to generate a spatial frequency spectrum, and the threshold determination unit 20 specifies Spectral power threshold, which is a threshold of the ratio of spatial low frequency power, is determined based on the number of reduced gradation bits, and the spatial frequency spectrum is acquired for each phase position by the original image pseudo contour region detection unit 30 to obtain the same phase position An area where the spatial low-frequency power ratio exceeds the spectrum power threshold is detected as a pseudo contour area. Since the threshold value determination unit 20 determines the spectrum power threshold value so as to decrease as the number of gradation reduction bits increases, according to the pseudo contour region detection device 1 of the first embodiment, the threshold determination unit 20 is appropriate according to the number of gradation reduction bits. A pseudo contour region can be detected.

  Moreover, it is preferable that the original image spatial frequency converter 10 converts the original image into the spatial frequency domain by performing multi-resolution analysis without decimation. In this case, it becomes possible to detect the pseudo contour region with high accuracy with the resolution accuracy (one pixel unit) of the original image.

  In addition, in order to function as the above-described pseudo contour region detection device 1, a computer can be suitably used. Such a computer can store a program describing processing contents for realizing each function of the pseudo contour region detection device 1. It can be realized by storing the program in a storage unit of the computer and reading and executing the program by the CPU of the computer.

  Next, a pseudo contour region detection apparatus according to the second embodiment will be described. In the pseudo contour region detection apparatus according to the second embodiment, a final pseudo contour region is detected from the pseudo contour region and the spatial frequency spectrum of both the original image and the gradation-reduced image. FIG. 3 is a block diagram illustrating a configuration example of the pseudo contour region detection apparatus according to the second embodiment. As illustrated in FIG. 3, the pseudo contour region detection device 2 includes an original image spatial frequency conversion unit 10, a threshold determination unit 20, an original image pseudo contour region detection unit 30, a tone reduction unit 40, and tone reduction. The image spatial frequency conversion unit 50, the gradation-reduced image pseudo contour region detection unit 60, and the total pseudo contour region detection unit 70 are provided.

  The original image spatial frequency conversion unit 10 converts the original image into a spatial frequency domain by a method having phase information, and generates a spatial frequency spectrum of the original image. Then, the generated spatial frequency spectrum of the original image is output to the original image pseudo contour region detection unit 30 and the total pseudo contour region detection unit 70. In order to detect the pseudo contour region with high resolution accuracy of the original image, it is preferable to perform multi-resolution analysis of the original image without decimation. When the original image spatial frequency conversion unit 10 performs multi-resolution analysis on the original image with decimation, the original image pseudo contour region detection unit 30 determines the size of the image block existing at the same phase position in each layer image. It is necessary to set the value according to the wavelet decomposition rank.

  Based on the number of gradation reduction bits set by the user, the threshold determination unit 20 determines a spectrum power threshold, which is a threshold of the ratio of spatial low frequency power, so that it decreases as the number of gradation reduction bits increases. Then, the determined spectral power threshold value is output to the original image pseudo contour region detection unit 30 and the gradation-reduced image pseudo contour region detection unit 60.

  The original image pseudo contour region detection unit 30 acquires the spatial frequency spectrum of the original image generated by the original image spatial frequency conversion unit 10 for each phase position, and the ratio of the spatial low frequency power at the same phase position is the threshold value determination unit 20. A region exceeding the spectrum power threshold determined by the above is detected as a pseudo contour region of the original image. However, unlike the first embodiment, the spatial frequency spectrum of the original image generated by the original image pseudo contour region detection unit 30 is not a pseudo contour region that is finally output by the pseudo contour region detection device 2, but a predetermined process described later. After performing the above, the final pseudo contour region is determined. The original image pseudo contour region detection unit 30 outputs the detected pseudo contour region of the original image to the total pseudo contour region detection unit 70.

  The gradation reduction unit 40 generates a gradation reduced image in which the gradation of the original image is reduced according to the number of gradation reduction bits designated by the user. Then, the generated gradation-reduced image is output to the gradation-reduced image spatial frequency conversion unit 50. For example, when 2-bit gradation reduction is performed from 10 bits (1024 gradations) to 8 bits (256 gradations), an image having an integer value obtained by dividing the gradation of the original image by 4 is used as the gradation-reduced image.

  The gradation reduced image spatial frequency conversion unit 50 converts the gradation reduced image generated by the gradation reduction unit 40 into a spatial frequency domain by a method having phase information, and generates a spatial frequency spectrum of the gradation reduced image. . Then, the generated spatial frequency spectrum of the gradation-reduced image is output to the gradation-reduced image pseudo contour region detection unit 60 and the total pseudo contour region detection unit 70. In order to detect the pseudo contour region with high resolution accuracy of the gradation-reduced image, it is preferable to perform multi-resolution analysis of the gradation-reduced image without decimation.

  The gradation-reduced image pseudo contour region detection unit 60 acquires, for each phase position, the spatial frequency spectrum of the gradation-reduced image generated by the gradation-reduced image spatial frequency conversion unit 50, and the spatial low-frequency power at the same phase position. A region where the ratio exceeds the spectrum power threshold determined by the threshold determination unit 20 is detected as a pseudo contour region of the gradation-reduced image. Then, the detected pseudo contour region of the gradation-reduced image is output to the total pseudo contour region detection unit 70.

  The total pseudo contour region detection unit 70 is a pseudo contour region of the original image detected by the original image pseudo contour region detection unit 30 and a pseudo contour of the gradation reduced image detected by the gradation reduced image pseudo contour region detection unit 60. For an area common to the areas, the spatial frequency spectrum of the original image generated by the original image spatial frequency conversion unit 10 is compared with the spatial frequency spectrum of the gradation reduced image generated by the gradation reduced image spatial frequency conversion unit 50. Then, a region where the spatial frequency spectrum of the gradation-reduced image is larger than the spatial frequency spectrum of the original image exceeds a predetermined threshold is detected as a final pseudo contour region. FIG. 4 is a block diagram illustrating a configuration example of the total pseudo contour region detection unit 70. As shown in FIG. 4, the total pseudo contour region detection unit 70 includes a common pseudo contour region detection unit 71 and a determination unit 72.

  The common pseudo contour region detection unit 71 is a pseudo contour region of the original image detected by the original image pseudo contour region detection unit 30 and a pseudo contour of the gradation reduced image detected by the gradation reduced image pseudo contour region detection unit 60. An area common to the areas is detected as a common pseudo contour area. Then, the detected common pseudo contour region is output to the determination unit 72. For example, when the pseudo contour region of the original image and the pseudo contour region of the gradation-reduced image are binary images in which the pseudo contour region is set to 1 and the other portions are set to 0, the logical product of both is calculated. Thus, the common pseudo contour region of the binary image can be detected.

  The determination unit 72 uses the spatial frequency spectrum of the original image generated by the original image spatial frequency conversion unit 10 and the tone reduction image spatial frequency conversion unit 50 for the common pseudo contour region detected by the common pseudo contour region detection unit 71. The spatial frequency spectra of the gradation-reduced images generated by the above are compared. The region where the spatial frequency spectrum of the gradation-reduced image is larger than the spatial frequency spectrum of the original image beyond a predetermined threshold is considered to be a pseudo contour due to the reduction of the gradation of the original image. The unit 72 detects the area as a final pseudo contour area.

  As described above, in the pseudo contour region detection device 2 according to the second embodiment, the original image spatial frequency conversion unit 10 converts the original image into the spatial frequency region to generate a spatial frequency spectrum, and the threshold determination unit 20 generates the spatial frequency spectrum. Spectral power threshold, which is a threshold of the ratio of spatial low frequency power, is determined so as to decrease as the key reduction bit number increases, and the pseudo contour region detection unit 30 acquires the spatial frequency spectrum for each phase position to obtain the same phase position. An area where the spatial low-frequency power ratio exceeds the spectrum power threshold is detected as a pseudo contour area. Further, the gradation reduction unit 40 generates a gradation reduction image of the original image according to the number of gradation reduction bits, and the gradation reduction image spatial frequency conversion unit 50 converts the gradation reduction image into a spatial frequency domain. Then, the spatial frequency spectrum of the gradation-reduced image is generated, and the spatial frequency spectrum of the gradation-reduced image is obtained for each phase position by the gradation-reduced image pseudo contour region detection unit 60, and the spatial low-frequency power at the same phase position is obtained. Is detected as a pseudo contour region of the gradation-reduced image. Then, the total pseudo contour region detection unit 70 calculates the spatial frequency spectrum of the original image and the spatial frequency spectrum of the tone reduced image for the region common to the pseudo contour region of the original image and the pseudo contour region of the tone reduced image. In comparison, a region where the spatial frequency spectrum of the gradation-reduced image is larger than the spatial frequency spectrum of the original image exceeding a predetermined threshold is detected as a final pseudo contour region. As described above, according to the pseudo contour area detection device 2 of the second embodiment, the final pseudo contour area is detected from both the pseudo contour areas of the original image and the gradation-reduced image. Compared with the detection apparatus 1, it becomes possible to detect a pseudo contour region with higher accuracy.

  Further, it is preferable that the original image spatial frequency conversion unit 10 and the tone reduction image spatial frequency conversion unit 50 perform multi-resolution analysis of the original image without decimation and convert the original image into the spatial frequency domain. In this case, it becomes possible to detect the pseudo contour region with high accuracy with the resolution accuracy (one pixel unit) of the original image.

  In addition, a computer can be preferably used to cause the above-described pseudo contour region detection device 2 to function, and such a computer can store a program describing processing contents for realizing each function of the pseudo contour region detection device 2. It can be realized by storing the program in a storage unit of the computer and reading and executing the program by the CPU of the computer.

  Each of the above embodiments has been described as a representative example, but it will be apparent to those skilled in the art that many changes and substitutions can be made within the spirit and scope of the invention. Therefore, the present invention should not be construed as being limited by the above-described embodiments, and various modifications and changes can be made without departing from the scope of the claims.

  As described above, according to the present invention, the pseudo contour region can be detected with high accuracy. Therefore, any use for detecting the pseudo contour region, or different quantization steps for the pseudo contour region and the non-pseudo contour region can be used. This is useful for applications such as requantization to reduce gradation.

DESCRIPTION OF SYMBOLS 1, 2 Pseudo contour area detection apparatus 10 Original image spatial frequency conversion part 20 Threshold determination part 30 Original image pseudo contour area detection part 40 Gradation reduction part 50 Gradation reduction image spatial frequency conversion part 60 Gradation reduction image pseudo contour area detection Unit 70 general pseudo contour region detection unit 71 common pseudo contour region detection unit 72 determination unit

Claims (5)

A pseudo contour region detection device that detects, as a pseudo contour region, an image that may generate a pseudo contour when the gradation of an original image is reduced,
An original image spatial frequency conversion unit that converts the original image into a spatial frequency domain and generates a spatial frequency spectrum of the original image;
A threshold value determining unit that determines a spectral power threshold value that is a threshold value of the spatial low-frequency power ratio based on the specified number of gradation reduction bits;
An original image pseudo contour region detection unit that acquires a spatial frequency spectrum of the original image for each phase position, and detects a region where a ratio of spatial low frequency power at the same phase position exceeds the spectrum power threshold as a pseudo contour region of the original image And comprising
The pseudo contour region detection apparatus, wherein the threshold value determination unit determines the spectrum power threshold value so as to decrease as the number of gradation reduction bits increases.   The pseudo contour region detection apparatus according to claim 1, wherein the original image spatial frequency conversion unit converts the original image into the spatial frequency region by performing multi-resolution analysis without decimation. A gradation reduction unit that generates a gradation-reduced image of the original image according to the number of gradation-reduction bits;
A gradation-reduced image spatial frequency converter that converts the gradation-reduced image into a spatial frequency domain and generates a spatial frequency spectrum of the gradation-reduced image;
Gradation reduction in which the spatial frequency spectrum of the gradation-reduced image is obtained for each phase position, and a region where the ratio of spatial low-frequency power at the same phase position exceeds the spectrum power threshold is detected as a pseudo contour region of the gradation-reduced image An image pseudo contour region detection unit;
Compare the spatial frequency spectrum of the original image and the spatial frequency spectrum of the gradation-reduced image for the area that is common to the pseudo-contour area of the original image and the pseudo-contour area of the gradation-reduced image, and reduce the gradation A comprehensive pseudo contour region detection unit that detects a region where the spatial frequency spectrum of the image exceeds a predetermined threshold value than the spatial frequency spectrum of the original image as a final pseudo contour region;
The pseudo contour region detection device according to claim 1, further comprising:   The pseudo contour region detection apparatus according to claim 3, wherein the tone-reduced image spatial frequency converting unit converts the tone-reduced image into the spatial frequency region by performing multiresolution analysis without decimation.   A program for causing a computer to function as the pseudo contour region detection device according to any one of claims 1 to 4.

Images