JP4629431B2 - Image processing apparatus and image processing method - Google Patents

Description

The present invention relates to an image processing apparatus, and relates to an image processing method, in particular, an image processing apparatus to enhance the contour component of the image data, and an image processing method.

  At present, the contour component of the image data is extracted when the image data is illustrated in the input image data. The filters used in that case are a Sobel filter as a primary differential filter and a Laplacian filter as a secondary differential filter. By these filters, the input image data is converted into image data from which contour components are extracted and output.

Here, the target pixel in the image data and the image unit data including the neighboring pixel are filtered to output the target pixel value, the neighboring pixel value is detected, and the filtered pixel value and the neighboring pixel value are detected. There is a technique for outputting image data from which a contour component is extracted by selecting any of (see, for example, Patent Document 1).
JP 2000-322569 A

  However, in the technique disclosed in Patent Document 1, since the input image data is directly filtered, cases where the target pixel value and the neighboring pixel value are simply different are extracted. Therefore, not only contour components but also essential unnecessary components such as fine texture components such as turf eyes and facial wrinkles are extracted from the input image data.

SUMMARY An advantage of some aspects of the invention is that it provides an image processing apparatus and an image processing method that enhance contour components more accurately .

In order to solve the above problems, an image processing apparatus having a subtraction circuit, a comparison circuit, a selection circuit, and a multiplication and addition circuit is provided.

In this image processing device, the subtraction circuit receives each pixel value of a pixel group including the target pixel in the image and a plurality of neighboring pixels located in the vicinity thereof, and receives the target pixel from the pixel value of each neighboring pixel in the pixel group. Are subtracted, the subtraction result is made an absolute value, and the difference value is output. The comparison circuit compares the output difference value with the threshold value stored in the threshold value memory. The selection circuit receives the pixel values of the pixel of interest and the neighboring pixels in units of pixel groups, and if the difference value is equal to or greater than the threshold based on the comparison result by the comparison circuit, the pixel of the neighboring pixel used for the calculation of the difference value The value is output as the selected pixel value, and when the difference value is smaller than the threshold value, the pixel value of the target pixel used for the calculation of the difference value is output as the selected pixel value. The multiplication and addition circuit refers to a coefficient memory in which a coefficient is stored for each pixel position in the pixel group, and each selected pixel based on the pixel value of the target pixel in the pixel group and each neighboring pixel in the pixel group Each value is multiplied by the coefficient in the coefficient memory corresponding to the position of the pixel of interest and each neighboring pixel in the pixel group, and the multiplication result for each pixel group is added and output.

In order to solve the above-mentioned problem, an image processing method in an image processing apparatus is provided. In this image processing method, the image processing apparatus executes the following processing.

The subtracting means of the image processing device accepts each pixel value of a pixel group including the target pixel in the image and a plurality of neighboring pixels located in the vicinity thereof, and calculates the pixel of the target pixel from the pixel value of each neighboring pixel in the pixel group Subtract the value, make the subtraction result an absolute value, and output it as a difference value. The comparison unit of the image processing apparatus compares the output difference value with the threshold value stored in the threshold value memory. The selection unit of the image processing apparatus is used to calculate the difference value when the pixel values of the target pixel and the neighboring pixels are input in units of pixel groups and the difference value is equal to or greater than the threshold value based on the comparison result by the comparison unit. If the difference value is smaller than the threshold value, the pixel value of the target pixel used for the calculation of the difference value is output as the selected pixel value. The multiplication / addition means of the image processing apparatus refers to a coefficient memory in which a coefficient is stored for each pixel position in the pixel group, and is based on the pixel value of the target pixel in the pixel group and each neighboring pixel in the pixel group. Each selected pixel value is multiplied by a coefficient in a coefficient memory corresponding to the position of the pixel of interest and each neighboring pixel in the pixel group, and the multiplication result in units of pixel groups is added and output.

According to the image processing apparatus, the contour component is more accurately emphasized in the input image.

Further, according to the above image processing method, the contour component is more accurately emphasized in the input image.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
First, the concept of the present invention will be described. FIG. 1 is a conceptual diagram of the present invention.
As shown in the figure, the image processing apparatus 10 includes terminals 11 and 12, a subtraction circuit 13, a threshold memory 14, a comparison circuit 15, a selection circuit 16, a coefficient memory 17, a multiplication and addition circuit 18, and a terminal 19. .

  Here, for the image data to be processed in which a plurality of pixels are arranged in the horizontal axis direction and the vertical axis direction, the image processing apparatus 10 uses one pixel in the image data as a target pixel and exists in the vicinity thereof. Image unit data having a pixel to be processed as a neighboring pixel is received. Then, the image processing apparatus 10 performs a predetermined process using the target pixel value of the target pixel and the neighboring pixel value of the neighboring pixel. The image unit data in the image data on which the predetermined processing is executed is arbitrarily set. Hereinafter, the image unit data in the image data will be described as 3 × 3 9 pixels. One pixel at the center of the nine pixels is set as a target pixel, and eight pixels other than the target pixel in the nine pixels are set as neighboring pixels. One pixel value is expressed by 8 bits. When the pixel value is close to 0, it is close to black, and when the pixel value is close to 255, it is close to white. In this image data, all pixels are set as target pixels in order from the upper left pixel to the lower right pixel, and an output target pixel value in which a contour component is extracted from the target pixel value is output. By doing so, the contour component is extracted and new image data is output.

  First, the threshold value memory 14 stores a threshold value that defines the magnitude of the difference value between the neighboring pixel value and the target pixel value. The coefficient memory 17 stores coefficients that are multiplied by the neighboring pixel value and the target pixel value, respectively. For example, when the Laplacian filter is stored in the coefficient memory 17, the coefficient for one target pixel value is -8, and the coefficient for eight neighboring pixel values is +1.

  Thereafter, the subtraction circuit 13 receives image unit data including one target pixel and eight neighboring pixels in the image data via the terminals 11 and 12. Then, the subtraction circuit 13 subtracts the target pixel value from the neighboring pixel values for the eight neighboring pixel values, sets the subtraction result to an absolute value, and outputs the result as eight difference values.

  Thereafter, the comparison circuit 15 compares the difference value with the threshold value for the eight difference values. Here, for example, when the difference value is 31 or less, if the neighboring pixel is an originally unnecessary component, if the difference value is 32 or more, the neighboring pixel becomes a contour component. Therefore, the threshold value is set to 32 in order to extract neighboring pixels having a difference value of 32 or more. Then, the selection circuit 16 to which the target pixel value and the neighboring pixel value are input, based on the comparison result by the comparison circuit 15 for the eight neighboring pixel values, selects the neighboring pixel value when the difference value is equal to or larger than the threshold. Output as pixel value. When the difference value is smaller than the threshold value, the selection circuit 16 replaces the neighboring pixel value with the target pixel value and outputs the selected neighboring pixel value. Here, when a certain neighboring pixel value is replaced with the target pixel value, the neighboring pixel value is not subjected to the filtering process to the target pixel value, so that the neighboring pixel is not subjected to the filtering process.

  Thereafter, the multiplication / addition circuit 18 multiplies the pixel value group composed of the selected neighboring pixel value and the target pixel value by nine coefficients corresponding to the position of each pixel in the image unit data. Then, the multiplication / addition circuit 18 adds all the multiplication results and outputs the result to the terminal 19 as an output target pixel value.

Next, an example of the operation of the image processing apparatus 10 in FIG. 1 will be described.
First, it is assumed that the threshold memory 14 stores the threshold “32”. Further, it is assumed that the coefficient memory 17 stores coefficients in the case of a Laplacian filter.

  Thereafter, the image unit data is received by the subtraction circuit 13. Then, the subtracting circuit 13 subtracts the pixel value of interest from the neighboring pixel values for the eight neighboring pixel values, makes the subtraction result an absolute value, and outputs it as eight difference values.

  Thereafter, the comparison circuit 15 compares the difference value with the threshold “32” for the eight difference values. Then, when the difference value is equal to or greater than the threshold value “32”, the selection circuit 16 outputs the neighboring pixel value as the selected neighboring pixel value based on the comparison result of the eight neighboring pixel values. When the difference value is smaller than the threshold value “32” by the selection circuit 16, the neighboring pixel value is replaced with the target pixel value and output as the selected neighboring pixel value.

  Thereafter, the multiplication and addition circuit 18 multiplies the pixel value group constituted by the selected neighboring pixel value and the target pixel value by the Laplacian filter. Then, all multiplication results are added by the multiplication / addition circuit 18 and output as an output target pixel value.

  In this way, the neighboring pixel value of the contour component whose difference value is equal to or greater than the threshold value is filtered, and the neighboring pixel value of the originally unnecessary component whose difference value is smaller than the threshold value is replaced with the target pixel value. Only the contour component of the image data can be extracted.

  The image unit data in the image data may be composed of 5 × 5 25 pixels and 3 × 5 15 pixels. In this case, one pixel at the center of the 25 pixels is the target pixel, and 24 pixels other than the target pixel in the 25 pixels are the neighboring pixels. In addition, one central pixel in the 15 pixels is a target pixel, and 14 pixels other than the target pixel in the 15 pixels are neighboring pixels.

Moreover, although the coefficient is multiplied to the selected neighboring pixel value, it can be multiplied to the difference value.
Next, specific contents of the embodiment of the present invention will be described. Here, the overall configuration will be described and described in detail with reference to FIG. FIG. 2 is a circuit diagram of the overall configuration of the image processing apparatus.

  As shown in the figure, the image processing apparatus 30 includes a terminal 31, terminals 32-1 to 32-8, arithmetic circuits 33-1 to 33-8, a multiplier circuit 36, an adder circuit 37, an offset value memory 38, and an arithmetic circuit 39. And terminal 40.

  First, the image processing apparatus 30 receives one target pixel value in the image unit data in the image data through the terminal 31, and receives eight neighboring pixel values through the terminals 32-1 to 32-8. Then, the multiplication circuit 36 performs a predetermined process such as a filter process on one pixel value of interest and outputs the result to the adder circuit 37. In addition, the arithmetic circuits 33-1 to 33-8 execute predetermined processing such as filter processing on the eight neighboring pixel values and output the result to the adding circuit 37.

  Thereafter, the adding circuit 37 adds all the processing results obtained by the arithmetic circuits 33-1 to 33-8 and the multiplying circuit 36. Then, the arithmetic circuit 39 calculates the offset value stored in the offset value memory 38 with respect to the addition result by the adder circuit 37 and outputs the result to the terminal 40 as the output target pixel value.

Next, the arithmetic circuits 33-1 to 33-8 in FIG. 2 will be specifically described, and the image processing apparatus 30 will be described in detail. FIG. 3 is a detailed circuit diagram of the image processing apparatus.
As shown in the figure, the image processing apparatus 30 includes terminals 31 and 32, a subtraction circuit 33a, a threshold memory 34, a comparison circuit 33b, a selection circuit 33c, a coefficient memory 35, multiplication circuits 33d and 36, an addition circuit 37, and an offset value memory. 38, an arithmetic circuit 39, and a terminal 40. Note that the subtraction circuit 33a, the comparison circuit 33b, the selection circuit 33c, and the multiplication circuit 33d may be one arithmetic circuit 33.

  Here, the image processing apparatus 30 has one pixel in the image data as a target pixel in the image data to be processed in which a plurality of pixels are arranged in the horizontal axis direction and the vertical axis direction, and exists in the vicinity thereof. To be received as neighboring pixels. Then, the image processing device 30 executes a predetermined process using the target pixel value of the target pixel and the neighboring pixel value of the neighboring pixel. The image unit data in the image data on which the predetermined processing is executed is arbitrarily set. Hereinafter, the image unit data in the image data will be described as 3 × 3 9 pixels. One pixel at the center of the nine pixels is set as a target pixel, and eight pixels other than the target pixel in the nine pixels are set as neighboring pixels. One pixel value is expressed by 8 bits. When the pixel value is close to 0, it is close to black, and when the pixel value is close to 255, it is close to white. In this image data, all pixels are set as target pixels in order from the upper left pixel to the lower right pixel, and an output target pixel value in which a contour component is extracted from the target pixel value is output. By doing so, the contour component is extracted and new image data is output.

  First, the threshold value memory 34 stores a threshold value that defines the magnitude of the difference value between the neighboring pixel value and the target pixel value. The coefficient memory 35 stores coefficients that are multiplied by the neighboring pixel value and the target pixel value, respectively. For example, when the Laplacian filter is stored in the coefficient memory 35, the coefficient for one target pixel value is -8, and the coefficient for eight neighboring pixel values is +1. Further, the offset value memory 38 stores an offset value for further adjusting the addition result by the addition circuit 37.

  Thereafter, the subtraction circuit 33a receives image unit data including one target pixel and eight neighboring pixels in the image data via the terminals 31 and 32. Then, the subtraction circuit 33a subtracts the target pixel value from the neighboring pixel values for the eight neighboring pixel values, sets the subtraction result to an absolute value, and outputs the result as eight difference values.

  Thereafter, the comparison circuit 33b compares the difference value with the threshold value for the eight difference values. Here, for example, when the difference value is 31 or less, if the neighboring pixel is an originally unnecessary component, if the difference value is 32 or more, the neighboring pixel becomes a contour component. Therefore, the threshold value is set to 32 in order to extract neighboring pixels having a difference value of 32 or more. The selection circuit 33c, to which the target pixel value and the neighboring pixel value are input, selects the neighboring pixel value based on the comparison result of the eight neighboring pixel values by the comparison circuit 33b when the difference value is equal to or larger than the threshold value. Output as pixel value. When the difference value is smaller than the threshold value, the selection circuit 33c replaces the neighboring pixel value with the target pixel value and outputs the selected neighboring pixel value. Here, when a certain neighboring pixel value is replaced with the target pixel value, the neighboring pixel value is not subjected to the multiplication with the target pixel value, and thus the neighboring pixel is not subjected to the filtering process.

  Thereafter, the multiplication circuit 33d multiplies the pixel value group constituted by the selected neighboring pixel values by eight coefficients corresponding to the positions of the respective pixels in the image unit data. Further, the multiplication circuit 36 multiplies the target pixel value by one coefficient corresponding to the position of each pixel in the image unit data. Then, the adder circuit 37 adds all the multiplication results obtained by the multiplier circuits 33d and 36.

  Thereafter, the arithmetic circuit 39 calculates an offset value for the addition result by the adder circuit 37 and outputs it to the terminal 40 as an output target pixel value. For example, the contour component of the image data has a large difference value and is extracted with the pixel value emphasized. Therefore, the pixels of the contour component are close to white and become a white contour component with respect to the black background component. As a countermeasure against this, the arithmetic circuit 39 inverts white and black or changes the density of the pixel.

Next, an example of the operation of the image processing apparatus 30 in FIG. 3 will be described.
First, it is assumed that the threshold memory 34 stores a threshold “32”. Further, it is assumed that the coefficient memory 35 stores coefficients in the case of a Laplacian filter. It is assumed that the offset value memory 38 stores an offset value that inverts white and black for the pixel as the addition result by the addition circuit 37.

  Thereafter, the image unit data is received by the subtraction circuit 33a. Then, the subtracting circuit 33a subtracts the pixel value of interest from the neighboring pixel values for the eight neighboring pixel values, makes the subtraction result an absolute value, and outputs it as eight difference values.

  Thereafter, the comparison circuit 33b compares the difference value with the threshold “32” for the eight difference values. Then, when the difference value is greater than or equal to the threshold value “32”, the selection circuit 33c outputs the neighboring pixel value as the selected neighboring pixel value based on the comparison result of the eight neighboring pixel values. When the difference value is smaller than the threshold value “32” by the selection circuit 33c, the neighboring pixel value is replaced with the target pixel value and output as the selected neighboring pixel value.

  Thereafter, the multiplication circuit 33d multiplies the pixel value group composed of the selected neighboring pixel values by the Laplacian filter. In addition, the multiplication circuit 36 multiplies the target pixel value and the Laplacian filter. Then, all the multiplication results by the multiplication circuits 33d and 36 are added by the addition circuit 37.

After that, the arithmetic circuit 39 calculates so as to invert white and black, and outputs the result as an output target pixel value.
Hereinafter, the threshold memory 34, the coefficient memory 35, and the offset value memory 38 will be described.

First, the threshold memory 34 will be described.
The threshold value memory 34 stores a threshold value that defines the magnitude of the difference value between the neighboring pixel value and the target pixel value. In other words, if it is equal to or greater than a certain difference value, it is regarded as a contour component, and the difference value is stored as a threshold value. For example, the threshold memory 34 stores a threshold for animation image data, a threshold for movie image data, a threshold for image data generated from a photograph, and the like. Here, since various threshold values can be set for the image data to be image-processed, various image data can be handled. Moreover, since the user who uses the image processing apparatus 30 can change the threshold value, the image processing result can be controlled based on experience.

Next, the coefficient memory 35 will be described. FIG. 4 is a diagram illustrating an example of the coefficient memory.
The coefficient memory 35 stores coefficients for multiplying the neighboring pixel value and the target pixel value, respectively. As shown in the figure, the coefficient memory 35 includes a filter name and nine coefficients. The filter name is a name of a filter used for image processing. The nine coefficients are a group of coefficients specific to the filter related to the filter name. For example, the A filter can extract a contour component, the B filter can extract a contour component in the horizontal axis direction, and the C filter can extract a contour component in the vertical axis direction. Here, since various image filters can be prepared assuming image data to be image-processed, various image data can be handled. Moreover, since the user who uses the image processing apparatus 30 can change each coefficient of the filter, the image processing result can be controlled based on experience.

Next, the offset value memory 38 will be described.
The offset value memory 38 stores an offset value for further adjusting the addition result by the addition circuit 37. Here, assuming image data to be image-processed, various offset values can be prepared, so that various image data can be handled. Moreover, since the user who uses the image processing apparatus 30 can change the offset value, the image processing result can be controlled based on experience.

  Next, the flow of image unit data by the image processing apparatus 30 will be described. FIG. 5 is a block diagram showing the flow of image unit data. For simplification of description, the arithmetic circuit 39 is omitted.

  First, the subtraction circuit 33a receives the image unit data 51. Then, the subtraction circuit 33a subtracts the target pixel value from the neighboring pixel values for the eight neighboring pixel values, sets the subtraction result to an absolute value, and outputs the result as eight difference values 52.

  After that, the comparison circuit 33 b compares the difference value with the threshold value 53 for the eight difference values 52. Then, based on the comparison result by the comparison circuit 33b for the eight neighboring pixel values, the selection circuit 33c outputs the neighboring pixel value as the selected neighboring pixel value when the difference value is greater than or equal to the threshold value 53. When the difference value is smaller than the threshold value 53, the selection circuit 33c replaces the neighboring pixel value with the target pixel value and outputs the selected neighboring pixel value. Then, new image unit data 54 is temporarily output.

  Thereafter, the multiplying circuit 33d multiplies the pixel value group constituted by the selected neighboring pixel value and the target pixel value by nine coefficients 55 corresponding to the positions of the respective pixels in the image unit data 54. Then, the addition circuit 37 adds all the multiplication results and outputs the result as an output target pixel value 56.

  The output target pixel value 56 is defined by the following equation, for example.

F (X) = 0: −ε ≦ X ≦ ε (2)
F (X) = X: X ≦ −ε X ≧ ε (3)
Here, in Expression (1), p` (0,0) is an output target pixel value, p (0,0) is a target pixel value, and p (a, b) is a neighboring pixel value. K (a, b) is a coefficient for a and b. These a and b indicate the positions of neighboring pixels from the target pixel, and are integers having values of −1, 0, and +1. Note that equation (1) describes only one pixel of interest. In Expressions (2) and (3), X is a difference value, and ε is a threshold value. Further, F (X) becomes X when the difference value is equal to or larger than the threshold value, and becomes 0 when the difference value is smaller than the threshold value.

  Next, the concept of processing in which the target pixel value is processed into the output target pixel value in the image processing device 30 will be described with reference to FIG. Note that the flowchart in FIG. 6 is executed when the image processing apparatus 30 receives image data. When the flowchart of FIG. 6 is executed, the following steps are executed. FIG. 6 is a flowchart showing processing of the image processing apparatus.

  Here, when the image unit data in the image data is 3 × 3 9 pixels, T = 3 and S = 3. A and b indicate positions of neighboring pixels from the target pixel, and are integers having values of −1, 0, and +1. When the flowchart is executed, the process shifts from the neighboring pixel (-1, + 1) to the neighboring pixel (+ 1, + 1), and from the neighboring pixel (-1,0) to the neighboring pixel (+1,0). Then, the process moves to the neighboring pixel (+1, -1) from the neighboring pixel (-1, -1). Further, the target pixel (0,0) has the target pixel value p (0,0), and the neighboring pixel (a, b) has the neighboring pixel value p (a, b).

[Step S11] The adder circuit 37 resets the output target pixel value p` (0, 0) to zero.
[Step S12] The comparison circuit 33b initially sets a = −1 and b = + 1. In other words, the comparison circuit 33b sets the neighboring pixel (a, b) to the leftmost and uppermost side.

  [Step S13] The comparison circuit 33b compares the difference value | p (a, b) -p (0,0) | with the threshold ε. If the difference value | p (a, b) −p (0,0) | is greater than or equal to the threshold ε, the process proceeds to step S14, and if smaller, the process proceeds to step S15.

  [Step S14] When the difference value | p (a, b) −p (0,0) | is equal to or larger than the threshold ε, the multiplication circuits 33d and 36 determine that the neighboring pixel value p (a, b) and the image unit data Is multiplied by a coefficient k (a, b) corresponding to the position of its neighboring pixel (a, b).

  [Step S15] When the difference value | p (a, b) −p (0,0) | is smaller than the threshold value ε, the multiplication circuits 33d and 36 recognize the pixel of interest replaced with the neighboring pixel value p (a, b). The value p (0,0) is multiplied by a coefficient k (a, b) corresponding to the position of the neighboring pixel (a, b) in the image unit data.

[Step S16] The adding circuit 37 adds all the multiplication results obtained by the multiplying circuits 33d and 36.
[Step S17] The comparison circuit 33b determines whether a is +1. In other words, the comparison circuit 33b determines whether or not the neighboring pixel (a, b) is the rightmost side. If it is +1, the process proceeds to step S19, and if it is not +1, the process proceeds to step S18.

[Step S18] If a is not +1, the comparison circuit 33b moves one pixel in the horizontal axis direction from the left side to the right side for the neighboring pixel (a, b).
[Step S19] When a is +1, the comparison circuit 33b resets a = -1. In other words, the comparison circuit 33b resets the neighboring pixel (a, b) to the leftmost side.

  [Step S20] The comparison circuit 33b determines whether b is -1. In other words, the comparison circuit 33b determines whether or not the neighboring pixel (a, b) is the lowermost side. If it is -1, the process proceeds to step S22, and if not -1, the process proceeds to step S21.

[Step S21] When b is not -1, the comparison circuit 33b moves one pixel in the vertical axis direction from the upper side to the lower side for the neighboring pixel (a, b).
[Step S <b> 22] When b is −1, the arithmetic circuit 39 calculates an offset value for the addition result by the adder circuit 37.

Note that the flowchart of FIG. 6 describes only one target pixel.
In this way, the neighboring pixel value of the contour component whose difference value is greater than or equal to the threshold value is filtered, and the neighboring pixel value of the originally unnecessary component whose difference value is smaller than the threshold value is replaced with the target pixel value. Only the contour component of the image data can be extracted.

  The image unit data in the image data may be composed of 5 × 5 25 pixels and 3 × 5 15 pixels. In this case, one pixel at the center of the 25 pixels is the target pixel, and 24 pixels other than the target pixel in the 25 pixels are the neighboring pixels. In addition, one central pixel in the 15 pixels is a target pixel, and 14 pixels other than the target pixel in the 15 pixels are neighboring pixels.

  Moreover, although the coefficient is multiplied to the selected neighboring pixel value, it can be multiplied to the difference value.

It is a conceptual diagram of this invention. 1 is a circuit diagram of an overall configuration of an image processing apparatus. It is a detailed circuit diagram of an image processing apparatus. It is a figure which shows the example of a coefficient memory. It is a block diagram which shows the flow of image unit data. It is a flowchart which shows the process of an image processing apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Image processing apparatus 11, 12, 19 Terminal 13 Subtraction circuit 14 Threshold memory 15 Comparison circuit 16 Selection circuit 17 Coefficient memory 18 Multiplication addition circuit

Claims (5)

Receiving each pixel value of a pixel group including the target pixel in the image and a plurality of neighboring pixels located in the vicinity thereof, subtracting the pixel value of the target pixel from the pixel value of each neighboring pixel in the pixel group, and a subtraction result A subtracting circuit that outputs an absolute value and a difference value;
A comparison circuit for comparing the difference value with a threshold value stored in a threshold memory;
When the pixel values of the pixel of interest and the neighboring pixels are input in units of the pixel group , and the difference value is equal to or greater than the threshold based on the comparison result by the comparison circuit, the neighborhood used for the calculation of the difference value and outputs the pixel value of the pixel as selection window pixel value if the difference value is smaller than the threshold value, a selection circuit for outputting a pixel value of the pixel of interest used in the calculation of the difference value as the selected pixel value,
Refer to a coefficient memory in which a coefficient is stored for each pixel position in the pixel group, and each selected pixel value based on the pixel value of the target pixel in the pixel group and the neighboring pixels in the pixel group Multiplication addition for multiplying each of the pixel of interest by the coefficient in the coefficient memory corresponding to the position of the pixel of interest and each of the neighboring pixels in the pixel group, and adding and outputting the multiplication result in units of the pixel group Circuit,
An image processing apparatus comprising: 2. The image processing apparatus according to claim 1, wherein each of the pixel groups includes an odd number of pixels of 3 or more in the horizontal axis direction and an odd number of pixels of 3 or more in the vertical axis direction. The image processing apparatus according to claim 1, wherein the threshold value memory stores a threshold value that is variable according to an input image. The coefficient memory stores a plurality of coefficient groups for emphasizing different types of contour components, and one of the plurality of coefficient groups is selectively input from the coefficient memory to the multiplication and addition circuit. The image processing apparatus according to claim 1. The subtracting means of the image processing device accepts each pixel value of a pixel group including the target pixel in the image and a plurality of neighboring pixels located in the vicinity thereof, and calculates the target pixel from the pixel value of each neighboring pixel in the pixel group. Subtracting the pixel value, making the subtraction result an absolute value, and outputting as a difference value;
A step of comparing the difference value and a threshold value stored in a threshold value memory by a comparison unit of the image processing apparatus ;
If the selection means of the image processing apparatus, each pixel value of the pixel of interest and the neighboring pixel is inputted in the pixel group unit, based on a comparison result by the comparison means, the difference value is greater than or equal to the threshold, the the pixel values of the neighboring pixels used in the calculation of the difference value and outputs it as selection window pixel value if the difference value is smaller than said threshold value, said selected pixel value of the pixel of interest used in the calculation of the difference value Outputting as pixel values ;
The multiplication and addition means of the image processing device refers to a coefficient memory in which a coefficient is stored for each pixel position in the pixel group, and the pixel value of the target pixel in the pixel group and the pixel value in the pixel group Each selected pixel value based on each neighboring pixel is multiplied by a coefficient in the coefficient memory corresponding to the position of the target pixel and each neighboring pixel in the pixel group, and the pixel group unit Adding and outputting the multiplication results ;
An image processing method comprising:

Images