JP2003174557A - Interpolation processing method, image processor, image forming device, image processing program and recording medium recorded with the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an interpolation processing method by which high image quality is maintained even after interpolation processing and the cost of an interpolation processing part is suppressed, and to provide an image processor, an image forming device, an image processing program, and a recording medium on which the image processing program, is recorded. <P>SOLUTION: Original image data are inputted, and the pixel values of prescribed peripheral pixels are extracted with respect to an interpolation pixel position. Gradation difference is calculated with an absolute value in each combination, and the calculated gradation difference is compared with a predetermined threshold to decide whether the gradation difference is bigger or smaller. Combinations with small gradation difference adjacent to one another are grouped on the basis of decision results, and the pattern of the grouped combinations is collated with predetermined table data. An interpolation operation control parameter predetermined in the collated pattern is set in an interpolation operation expression to calculate the pixel value of the interpolation pixel. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an interpolation processing method, an image processing apparatus, an image forming apparatus, an image processing program, and an interpolation processing method for obtaining a pixel value of an interpolation pixel by performing an interpolation process using pixel values of peripheral pixels. Recording medium

[0002]

2. Description of the Related Art Conventionally, there have been methods for interpolating images. Depending on the application, simple interpolation is performed to reduce the processing cost, and smooth interpolation is performed by relatively simple arithmetic processing. , There is one that tries to maintain the reproducibility of an image to some extent by more complicated arithmetic processing. There are three basic interpolation processing methods: Nearest Neighbor (hereinafter referred to as "NN"), Bi-liner (hereinafter referred to as "BL"), and Bi-Cubic convolution (hereinafter referred to as "BC"). ) And the like.

23 to 25 show examples of the above-mentioned three interpolation processing methods. In these figures, the pixels in the shadow portion represent existing pixels, and the pixels in the white background portion represent interpolated pixels.

FIG. 23 is a diagram showing an example of interpolation processing by the NN. NN sets the pixel value of the existing pixel closest to the interpolation pixel position as the pixel value of the interpolation pixel. In FIG. 23, among the existing pixels around the interpolation pixel, the pixel value of the pixel at the right, lower right, or just below position becomes the pixel value of the interpolation pixel. The NN is the simplest interpolation processing method and has a small processing amount and can be processed at high speed. However, the image quality is deteriorated due to jaggies in which curved lines and diagonal straight lines are jagged like stairs. There is.

FIG. 24 is a diagram showing an example of interpolation processing by BL. BL obtains the average of the values weighted in proportion to the distances of the four existing pixels around the interpolation pixel, and uses that average as the pixel value of the interpolation pixel. The amount of processing is not so large and it is excellent in that a smooth image can be obtained, but since the edge portion is also smoothed, there is a problem that the edge becomes dull and blurred.

FIG. 25 is a diagram showing an example of interpolation processing by BC. The BC performs the interpolation calculation with a total of 16 pixels, including the existing 4 pixels around the interpolated pixel and the existing 12 pixels surrounding them. In FIG. 25, the center pixel is an interpolation pixel, and the outermost shadow portion pixel corresponds to 12 existing pixels. BC is excellent in that it can obtain a smooth image to some extent and prevents blunting of edges, but since the operation is the most complicated among these three methods, the circuit scale becomes large and interpolation There is a problem that the processing cost is high.

There is also a method of changing the interpolation algorithm according to the scaling factor of enlargement or reduction, such as the image processing method described in Japanese Patent Laid-Open No. 2000-188689.
Specifically, in the enlargement / reduction processing, BC is used near the magnification where moire is likely to occur, and NN or BL is used near the other magnifications. By using BC, it is possible to suppress the occurrence of moire, and by using NN or BL, the processing time can be shortened, so that the average processing capacity per unit time does not decrease so much as a whole, It is possible to suppress deterioration of image quality due to occurrence of moire.

[0008]

[Problems to be Solved by the Invention]
Since there are the above-mentioned problems in the two interpolation processing methods,
The image quality of the original image cannot be sufficiently maintained depending on the magnification at the time of performing the interpolation processing or the type of the image such as a character, a photograph, a line drawing, or CG (Computer Graphics). That is, since the processing method is too simple, a large amount of information is lost and jaggies occur, but sufficient improvement processing is not performed on such images,
Even if the circuit scale is increased to complicate the processing method, there is a problem that a high image quality corresponding to the circuit scale cannot be obtained. Further, even if the three interpolation processing methods are properly used as in the image processing method described in Japanese Patent Laid-Open No. 2000-188689, the problem itself of the interpolation processing that each has is not solved, and therefore, sufficient image quality is achieved. Can't get

An object of the present invention is to provide an interpolation processing method, an image processing apparatus, which can maintain high image quality even after the interpolation processing and can suppress the cost of the interpolation processing unit.
An object of the present invention is to provide an image forming apparatus, an image processing program, and a recording medium recording the same.

[0010]

According to the present invention, in an interpolation processing method for obtaining a pixel value of an interpolated pixel by performing an interpolating process using the pixel value of a peripheral pixel, the gradation of the pixel value between the peripheral pixels at the interpolated pixel position. The difference is determined, the gradation difference is determined by comparing the gradation difference with a predetermined threshold value, and when the gradation difference is larger than the threshold value, the pixel value of the interpolation pixel is the pixel value of the nearest neighbor pixel among the peripheral pixels. Weighting to get closer to
When the gradation difference is smaller than the threshold value, the pixel value of the interpolated pixel is obtained using an interpolation calculation control parameter for weighting so that the pixel value of the interpolated pixel is linearly interpolated by the pixel value of the peripheral pixel. This is an interpolation processing method.

According to the present invention, when the gradation difference is larger than the threshold value, weighting is performed so that the pixel value of the interpolation pixel approaches the pixel value of the nearest neighbor pixel among the peripheral pixels, and when the gradation difference is smaller than the threshold value. , The pixel value of the interpolated pixel is obtained by using the interpolation calculation control parameter for weighting so that the pixel value of the interpolated pixel is linearly interpolated by the pixel value of the peripheral pixel, so that the interpolated weighting is different depending on the magnitude of the gradation difference. A calculation control parameter can be used, and sharpness can be preserved for an edge portion having a large gradation difference, while high image quality can be maintained by smoothly interpolating a flat portion having a small gradation difference.

Further, the present invention is characterized in that all the combinations that select two pixels from the peripheral pixels included in a predetermined peripheral pixel area are calculated, and the gradation difference is calculated for each combination. .

According to the present invention, since the gradation difference is obtained for each combination, it is possible to precisely discriminate which is the edge portion and which is the flat portion in the peripheral pixel area.

Further, according to the present invention, the combinations are grouped in a peripheral pixel area based on the result of the determination, and the patterns of the grouped combinations are collated with patterns set as table data in advance. Characterize.

According to the present invention, the patterns of the grouped combinations are collated with the patterns set in advance as table data, so that the interpolation calculation control parameter for obtaining the pixel value of the interpolation pixel can be uniquely determined. It is possible to realize the optimum interpolation processing for the edge portion or the flat portion which is finely distinguished. Further, since the table data is prepared in advance, the arithmetic processing can be simplified and the cost of the interpolation processing unit can be suppressed.

According to the present invention, in the collation, if the pattern of the combination does not match the pattern of the table data, the pattern of the combination is rotated by 90 degrees and the interpolation is performed according to the number of rotations until the pattern matches. It is characterized in that the setting order of the arithmetic control parameters is changed.

According to the present invention, the setting order of the interpolation calculation control parameters is changed according to the number of revolutions until they coincide with each other, so that it is not necessary to prepare an extra table data pattern, and the table data pattern is stored. The memory usage can be reduced.

Further, according to the present invention, in the interpolation processing method for interpolating using the pixel values of the peripheral pixels to obtain the pixel values of the interpolating pixels, the average value of the pixel values of the peripheral pixels at the interpolating pixel position is obtained to obtain the interpolating pixel values. While setting as a temporary pixel value,
A distance between the interpolation pixel position and the peripheral pixel and a gradation difference between the provisional pixel value and the pixel value of the peripheral pixel are obtained, and a non-linear function is used for the provisional pixel value, the distance, and the gradation difference. This is an interpolation processing method characterized in that the pixel value of the interpolation pixel is obtained.

According to the present invention, since the pixel value of the interpolated pixel is obtained by using the non-linear function for the provisional pixel value, the distance and the gradation difference, the sharpness is preserved for the edge portion having the large gradation difference. By smoothly interpolating a flat portion having a small gradation difference, high image quality can be maintained.

Further, the present invention is characterized in that the distance is a Euclidean distance between the interpolation pixel position and the peripheral pixel.

According to the present invention, the distance is the Euclidean distance between the interpolated pixel position and the peripheral pixel, so that which of the peripheral pixels is closest to the interpolated pixel position can be easily determined. . Moreover, since the calculation method is not complicated, the configuration of the device that implements the interpolation processing can be simplified, and the cost of the interpolation processing unit can be suppressed.

Further, the present invention is characterized in that the gradation difference is an absolute value of a difference between the provisional pixel value and the pixel values of the peripheral pixels.

According to the present invention, the gradation difference is the absolute value of the difference between the provisional pixel value and the pixel value of the peripheral pixels, so that which of the peripheral pixels is closest to the provisional pixel value is determined. You can decide. Moreover, since the calculation method is not complicated,
It is possible to simplify the configuration of the device that implements the interpolation processing and reduce the cost of the interpolation processing unit.

Further, the present invention is characterized in that a membership function is used as the non-linear function.

According to the present invention, since the membership function is used as the non-linear function, the vague sense of distance and the vague gradation smoothness can be digitized and evaluated, and the fuzzy rule determined in advance from these values. The pixel value of a valid interpolation pixel can be obtained by obtaining the conformity to

Further, the present invention is characterized in that the gradation difference is obtained using a luminance value and a pixel value representing a color component.

According to the present invention, since the gradation difference is obtained by using the brightness value and the pixel value representing the color component, there is no gradation difference of the brightness value but the color is different especially in the interpolation processing of the color image. It is possible to smoothly interpolate a flat portion while preserving sharpness even for a color edge portion visually recognized as an edge by human eyes.

Further, according to the present invention, in the image processing apparatus including an interpolation processing section for performing interpolation processing using pixel values of peripheral pixels to obtain pixel values of the interpolated pixels, A peripheral pixel value extracting means for extracting a value, a pixel selecting means for obtaining all combinations for selecting two pixels from the peripheral pixels included in a predetermined peripheral pixel area, and a gradation difference for each combination. ,
A tone difference determining means for comparing the tone difference with a predetermined threshold value to determine a tone difference; and a grouping means for grouping the combination in the peripheral pixel region based on the result of the determination, Pattern matching means for matching the grouped combination patterns with a pattern set in advance as table data, and an interpolation calculation control parameter set in advance for each of the table data is set in the interpolation calculation formula based on the result of the matching. An image processing apparatus comprising: an interpolation calculation control parameter setting unit; and an interpolation calculation processing unit that performs an interpolation calculation process using the interpolation calculation control parameter.

According to the present invention, the interpolation processing unit extracts peripheral pixel value extracting means for extracting pixel values of peripheral pixels at the interpolation pixel position and two pixels from the peripheral pixels included in the predetermined peripheral pixel area. Pixel selection means for determining all combinations to be selected, gradation difference determination means for determining a gradation difference for each combination, comparing the gradation difference with a predetermined threshold value to determine the gradation difference, and a combination of peripheral pixels Grouping means for grouping in the area based on the result of the judgment, pattern matching means for matching the grouped combination pattern with the pattern set in advance as table data, and for each table data based on the result of the matching Interpolation calculation control parameter setting means for setting the interpolation calculation control parameter predetermined in Since it has an interpolation calculation processing means for performing an interpolation calculation process by using it, it is possible to use an interpolation calculation control parameter for weighting differently depending on the magnitude of the gradation difference, and to preserve the sharpness for the edge part with a large gradation difference. On the other hand, high image quality can be maintained by smoothly interpolating the flat portion having a small gradation difference.

Further, according to the present invention, in the image processing apparatus having the interpolation processing unit for obtaining the pixel value of the interpolation pixel by using the pixel value of the peripheral pixel, the interpolation processing unit extracts the pixel value of the peripheral pixel at the interpolation pixel position. Peripheral pixel value extraction means, provisional pixel value setting means for obtaining the average value of the pixel values of the peripheral pixels at the interpolation pixel position and setting it as the provisional pixel value of the interpolation pixel, and obtaining the distance between the interpolation pixel position and each peripheral pixel. Distance calculating means, a gradation difference calculating means for calculating a gradation difference between the provisional pixel value and the pixel value of each peripheral pixel, and the calculated distance value as an input value,
First antecedent part membership value calculating means for outputting as a numerical value the degree of closeness of the distance between the interpolation pixel position and each peripheral pixel, and the calculated gradation difference value as an input value, and the temporary pixel Output value of the second antecedent part membership value calculating means for calculating the degree of smallness of the gradation value between the value and the pixel value of each peripheral pixel as a numerical value, and the first antecedent part membership value calculating means And the output value of the second antecedent part membership value calculation means are used as input values, and the interpolated pixel is calculated using the consequent part membership value calculation means and the output value of the consequent part membership value calculation means. And an interpolation calculation processing unit for obtaining the pixel value of.

According to the present invention, the interpolation processing section obtains the peripheral pixel value extraction means for extracting the pixel values of the peripheral pixels at the interpolation pixel position, and the average value of the pixel values of the peripheral pixels at the interpolation pixel position to obtain the interpolation pixel. Provisional pixel value setting means for setting the provisional pixel value, distance calculating means for determining the distance between the interpolated pixel position and each peripheral pixel, and gradation for determining the gradation difference between the provisional pixel value and the pixel value of each peripheral pixel. A difference calculating means, a first antecedent part membership value calculating means that outputs the calculated distance value as an input value, and outputs the degree of closeness of the distance between the interpolated pixel position and each peripheral pixel as a numerical value; The input value is the gradation difference value
Of the second antecedent part membership value calculating means and the first antecedent part membership value calculating means for calculating the degree of gradation value smallness between the provisional pixel value and the pixel value of each peripheral pixel as a numerical value. Using the consequent part membership value calculating means that uses the logical sum of the output value and the output value of the second antecedent part membership value calculating means and the output value of the consequent part membership value calculating means Since it has an interpolation calculation processing unit for obtaining the pixel value of the interpolated pixel, sharpness is preserved for an edge portion with a large gradation difference, while smooth interpolation is performed for a flat portion with a small gradation difference to obtain high image quality. Can be held.

Further, the present invention is characterized in that the gradation difference is obtained by using a pixel value representing a luminance value and a color component.

According to the present invention, since the gradation difference is obtained using the brightness value and the pixel value representing the color component, there is no gradation difference in the brightness value but the color is different particularly in the interpolation processing of a color image. Moreover, it is possible to smoothly interpolate the flat portion while preserving the sharpness even for the color edge portion visually recognized as an edge by human eyes.

The present invention is also an image forming apparatus including the image processing device, an image input device, and an image output device.

According to the present invention, since the image processing device, the image input device, and the image output device are provided, the flat portion having a small gradation difference is preserved while the sharpness is preserved for the edge portion having a large gradation difference. The smoothly interpolated image can be used for various purposes.

The present invention is also an image processing program for causing a computer to execute the interpolation processing method.
According to the present invention, it is possible to cause a computer to execute the interpolation processing.

Further, the present invention is a computer-readable recording medium in which the image processing program is recorded.
According to the present invention, it is possible to cause a computer to execute the interpolation processing.

[0038]

1 is a block diagram showing the configuration of an image forming apparatus 2 having an image processing apparatus 1 according to an embodiment of the present invention. The image processing apparatus 1 includes an A / D (analog / digital) conversion unit 11 and a shading correction unit 1.
2, input tone correction unit 13, region separation processing unit 14, color space conversion unit 15, spatial filter processing unit 16, interpolation processing unit 1
7, a color correction unit 18, a black generation undercolor removal unit 19, an output gradation correction unit 110, and a gradation reproduction processing unit 111. The image processing device 1 includes an image input device 21 and an image output device 2.
An external input interface unit such as an interface unit as a network printer via Ethernet (registered trademark) and an interface unit of a telephone line for transmitting and receiving a fax, for connecting the image forming apparatus 2 to which the image forming apparatus 2 is connected to the outside. The digital multi-function peripheral 3 is configured as a whole including the 31 and the external input / output interface 32.

The image input device 21 is, for example, a CCD (Ch
arge Coupled Device)
The CCD reflects the reflected light image from the original in RGB (R: red,
G: green, B: blue) are read as analog signals and input to the image processing apparatus 1.

The analog signal read by the image input device 21 is transferred to the A / D converter 1 in the image processing device 1.
1, a shading correction unit 12, an input gradation correction unit 13,
The area separation processing unit 14, the color space conversion unit 15, the spatial filter processing unit 16, the interpolation processing unit 17, the color correction unit 18, the black generation undercolor removal unit 19, the output gradation correction unit 110, and the gradation reproduction processing unit 111. Sent in order, CMYK (C: cyan, M:
It is output to the image output device 22 as a digital signal of magenta, Y: yellow, and K: black).

The A / D converter 11 converts the RGB analog signals read by the image input device 21 into digital signals. The shading correction unit 12 is the A / D conversion unit 1.
With respect to the RGB signal sent from the No. 1, processing for removing various distortions generated in the illumination system, the imaging system, and the imaging system of the image input device 21 is performed. The input gradation correction unit 13 is an RG whose various distortions have been removed by the shading correction unit 12.
For the signal based on the reflectance of B, color balance is adjusted, and at the same time, a signal such as a density signal is converted into a signal that can be easily handled by the image processing system adopted in the image processing apparatus 1. The area separation processing unit 14 uses the RGB signals to
Each pixel in the input image data is separated into a character area, a halftone dot area, and a photographic area, and based on the separation result, an area identification signal indicating to which area the pixel belongs, a spatial filter processing unit 16, The input signal output from the input tone correction processing unit 13 is directly output to the color space conversion unit 15 while being output to the black generation undercolor removal unit 19 and the tone reproduction processing unit 111.

The color space conversion unit 15 converts the RGB image data into an LCC (L: luminance, C: chromaticity) color space, and outputs the converted data to the spatial filter processing unit 16. The spatial filter processing unit 16 performs spatial filter processing by a digital filter on the image data of the LCC signal input from the color space conversion unit 15 based on the region identification signal, and corrects the spatial frequency characteristic to output the image. To prevent blurring and deterioration of graininess. For example, the area separated into characters by the area separation processing unit 14 has a large amount of high frequency emphasis due to the sharpness emphasis processing in the spatial filter processing by the spatial filter processing unit 16 in order to particularly improve the reproducibility of black characters or color characters. Become. Further, the area separated into halftone dots by the area separation processing unit 14 is subjected to low-pass filter processing in the spatial filter processing unit 16 in order to remove the input halftone dot component. The spatially filtered image data is output to the interpolation processing unit 17.

The interpolation processing unit 17 performs interpolation calculation processing on the spatially filtered image data using the pixel values of the peripheral pixels to obtain the pixel values of the interpolated pixels. Then, the image data after the interpolation processing is output to the color correction unit 18. Details will be described later.

The color correction unit 18 converts the LCC image data into C
The color is converted into the MY color space, the color is corrected in accordance with the image output device 22, and the corrected data is output to the black generation lower color removing unit 19. The black generation and undercolor removal unit 19 generates black for generating a K signal from the CMY three color signals after color correction;
Subtract the K signal obtained from black generation from the original CMY signal,
The process of generating a new CMY signal is performed, and the CMY signal 3
The color signals are converted into CMYK four-color signals. As an example of black generation processing, there is a method of generating black by skeleton black which is a general method. In this method, the input / output characteristic of the skeleton curve is y = f (x), the input data is C, M, Y, the output data is C ′, M ′,
If Y ′ and K ′ and the UCR (Under Color Removal) rate is α (0 <α <1), the black generation undercolor removal process can be expressed by the following equation.

[0045]

[Equation 1]

The gradation reproduction processing unit 111, like the spatial filter processing unit 16, performs a predetermined process on the image data of the CMYK signals based on the area identification signal. For example, with respect to the area separated into characters by the area separation processing unit 14, in order to enhance the reproducibility of black characters or color characters, the gradation reproduction processing unit 111 uses a high-resolution screen suitable for reproducing high frequencies. The binarization processing or the multi-value processing of is selected. Further, regarding the area separated into halftone dots by the area separation processing unit 14, the output tone correction unit 110
Then, after performing an output gradation correction process for converting a signal such as a density signal into a halftone dot area ratio which is a characteristic value of the image output device 22, a gradation reproduction processing unit 111 finally separates the image into pixels. Then, gradation reproduction processing is performed to reproduce each gradation, that is, halftone generation is performed. Regarding the areas separated into photographs by the area separation processing unit 14,
Binarization or multi-valued processing on the screen with emphasis on gradation reproducibility is selected.

The image data that has been subjected to each of the above processes is temporarily stored in the storage means, read at a predetermined timing, and input to the image output device 22. Image output device 22
Is to output image data to a recording medium such as paper.
For example, an image output device 22 using an electrophotographic method or an inkjet method is used, but the invention is not particularly limited to this.

The interpolation processing section 17 in the present invention is
Although the processing is performed after the processing of the spatial filter processing unit 16, after the input gradation correction unit 13, the gradation reproduction processing unit 11 such as after the color correction unit 18 and the black generation undercolor removal unit 19.
If it is 1 or earlier, the processing may be performed at any timing.

[First Embodiment] The first embodiment of the present invention will be described below. FIG. 2 is a block diagram showing the configuration of the interpolation processing unit 17 included in the image processing apparatus 1 of the present invention. The interpolation processing unit 17 includes a peripheral pixel value extraction unit 51 and a pixel selection unit 5.
2, gradation difference determination means 53, grouping means 54, pattern matching means 55, interpolation calculation control parameter setting means 56
And an interpolation calculation processing means 57.

The peripheral pixel value extraction means 51 extracts a pixel value for each peripheral pixel at the interpolation pixel position from the image data filtered by the spatial filter processing section 16.
The pixel value is a value including a brightness value and a color component. This extraction method will be described with reference to FIG. Figure 3
(A) is an interpolated pixel position P and its peripheral pixels P1 to P
FIG. For example, when four pixels are used as peripheral pixels, as shown in the figure, the interpolation pixel position P
The pixel P1 located at the upper left of is extracted, and four peripheral pixels P1, P2, P3, and P4 indicated by black circles surrounding the interpolation pixel position P in the pixel region of 8 neighborhoods surrounding P1 are extracted, and Extract the pixel value. As a result, the pixel values of the peripheral pixels can be uniformly extracted. The area including the peripheral pixels P1, P2, P3, and P4 is referred to as a peripheral pixel area.

The pixel selecting means 52 selects all the combinations for selecting two pixels from the peripheral pixels included in the predetermined peripheral pixel area. Figure 3 shows how to select
An explanation will be given using (b). FIG. 3B is a diagram showing all combinations for selecting two pixels. As shown in the figure, 2 selected from the pixels included in the peripheral pixel area
Combination of two pixels, ie, 2 sets of each pixel vertically, 2 horizontally
A total of 6 combinations of two sets and two diagonal sets are obtained as all combinations.

The gradation difference determining means 53 determines the gradation difference for each selected combination and compares it with a predetermined threshold value to judge the gradation difference. As a result, it is possible to minutely discriminate which is the edge portion and which is the flat portion in the peripheral pixel region. Grouping means 54
For grouping the combinations based on the result of the gradation difference determined in the peripheral pixel area to perform grouping. The pattern matching means 55 matches the grouped combination patterns with the patterns set in advance as table data. At that time, if no matching table data is found in one collation, the pattern of the combination is rotated by 90 degrees clockwise or counterclockwise up to 3 times until the table data is collated. Do up to 4 times. As a result, it is not necessary to prepare an extra table data pattern, and the amount of memory used to store the table data pattern can be reduced. Actually, since the order of determining the gradation difference is determined, 1 is set if the gradation difference is larger than the threshold value, and 1 if it is smaller than the threshold value.
Be sure to output bits. Then, the bits are arranged in the determined order to create a 6-bit pattern, and the result of the determination grouped by the pattern matching means 55 and the table data, that is, the 6-bit pattern and the 6-bit of the predetermined table data. Match the index. As a result, the interpolation calculation control parameter for obtaining the pixel value of the interpolation pixel can be uniquely determined, and the optimum interpolation processing can be realized for the edge portion or the flat portion that is finely distinguished. Further, since the table data is prepared in advance, the arithmetic processing can be simplified and the cost of the interpolation processing unit 17 can be suppressed.

The interpolation calculation control parameter setting means 56 is
Based on the comparison result, the interpolation calculation control parameter set in advance for each table data is set in the interpolation calculation formula. When setting, if rotation is performed during pattern matching, it is necessary to shift the order of the interpolation calculation control parameters by the number of times. The interpolation calculation processing means 57 performs interpolation calculation processing using the interpolation calculation control parameter set by the interpolation calculation control parameter setting means 56. Finally, the output image data after the interpolation processing is output to the output gradation correction unit 110. As described above, it is possible to use the interpolation calculation control parameter that weights differently depending on the magnitude of the gradation difference, and the sharpness is preserved for the edge portion where the gradation difference is large, while the smoothness is maintained for the flat portion where the gradation difference is small. Can be interpolated to.

Next, the interpolation processing in the image processing apparatus 1 will be specifically described by using a certain interpolation pixel position P.
FIG. 4 is a diagram showing a part of the input original image data. In the figure, data of 4 × 4 16 pixels is shown. Here, each data is 8-bit data and has a value of 0 to 255. For the interpolated pixel position P, the pixel values of the four black circle pixels are extracted as the pixel values of the peripheral pixels. As shown in the figure, the brightness value is 140 for P1 and P
2 is 132, P3 is 201, and P4 is 151.

Next, all combinations for selecting two pixels from the peripheral pixels included in the predetermined peripheral pixel area are obtained. FIG. 5 is a diagram showing all combinations for selecting two pixels from the peripheral pixels included in the peripheral pixel area of the interpolation pixel position P. From the extracted peripheral pixels, the gradation difference is calculated as an absolute value for all combinations that select two pixels, that is, for each combination. The gradation difference is 8, 50, 61, 19, 69, 11 in the order of numbers from to. Then, it is determined whether the gradation difference is larger or smaller than the threshold value. Here, a value of 20 is set as the threshold value. Note that this value is an experimentally set value, and the present invention is not limited to this value. When the gradation difference is determined by comparing the gradation difference with the threshold value, it means that the gradation difference of and of the combination is large and the gradation difference of and is small.

Next, based on the result of the judgment, the adjacent ones are grouped in a combination having a small gradation difference, thereby performing grouping based on the gradation difference. FIG. 6 is a diagram showing a result of grouping in the peripheral pixel area of the interpolation pixel position P. The pattern of grouped combinations is collated with predetermined table data. Here, in order to perform the matching, the combination pattern is compared with the table data of the pattern. FIG. 7 is a diagram showing pattern table data. In FIG. 7A, each pattern is shown by a black circle, a solid line, and a dotted line, and this solid line represents a combination in which the gradation difference is less than or equal to a predetermined threshold value, that is, a combination in which the gradation difference is small. FIG. 7B shows each local area in each pattern by a group number, and the group numbers are given in order from the smallest gradation difference. As shown in FIG. 6, the grouped combination pattern is divided into two regions, a region with a large gradation difference and a region with a small gradation difference, and is below the diagonal line with a diagonal line connecting P1 and P4 as a boundary. The region on the side has a large gradation difference, and the region above the diagonal line has a small gradation difference. Therefore, it can be seen that when the pattern of the combination is rotated three times clockwise, it coincides with the pattern 9. .

Next, the interpolation calculation control parameters set in advance according to the collated pattern are set in the interpolation calculation formula. FIG. 8 is a diagram showing interpolation calculation control parameters determined corresponding to each pattern of table data. The group number of each local area in the individual pattern of FIG. 7B corresponds to the branch number of the pattern number of FIG. Since the combination pattern rotates three times clockwise, it is necessary to shift the order of the interpolation calculation control parameters. For example, in the case of the parameter α, α → γ (1 rotation) → δ (2 rotations) → β depending on the rotation speed.
Since it is necessary to shift the order to (3 rotations), the combination patterns rotated clockwise three times are α ′ → β,
It is necessary to shift to β ′ → δ, γ ′ → α, δ ′ → γ (α ′, β ′, γ ′ and δ ′ represent interpolation calculation control parameters before rotation).

In the case of pattern 9, the interpolation calculation control parameter is set as follows. The interpolation calculation control parameter is β = α ′ = 1 + r / 12. From the mathematical expression of r in FIG. 8, r when β is r = {u−x} · {(y +
1) -v}, substituting this results in β = 1 +
It becomes {u−x} · {(y + 1) −v} / 12. Also δ
= Β ′ = β = 1 + r / 12, and substituting the mathematical expression of r in FIG. 8, δ = 1 + {u−x} · {v−y} / 12. Further, α = γ ′ = β = 1 + r / 12, and r in FIG.
Substituting the mathematical expression of, α = 1 + {(x + 1) -u}.
It becomes {(y + 1) -v} / 12. Also, γ = δ ′ = 0. In addition, in 9-2 (“group 2”) in which the interpolation calculation is performed using one nearest neighbor pixel, similarly substituting mathematical expressions, β = δ = α = 0, γ = 1 / {u− x} ・
It becomes {v−y}.

[0059]

[Equation 2]

Since the interpolation pixel position P belongs to the group 1, the interpolation calculation control parameter of the group 1 is used in the interpolation calculation. When the interpolated pixel position P belongs to the group 2, the interpolation calculation control parameter of the group 2 is used.

When the position of the interpolation pixel is P (u, v) and the position of the upper left peripheral pixel is P1 (x, y), the interpolation calculation formula of the present invention is expressed as follows. Note that x = [u], y
= [V] ([] represents integerization by rounding off after the decimal point). P (u, v) = {(x + 1) -u} * {(y + 1) -v} * P1 (x, y) * [alpha] + {u-x} * {(y + 1) -v} * P2 (x + 1, y) · β + {(x + 1) −u} · {v−y} · P3 (x, y + 1) · γ + {u−x} · {v−y} · P4 (x + 1, y + 1) · δ ... ( 4)

The interpolated pixel position P is P (u, v) = P (2
0.8, 16.3), x = 20 and y = 16. Substituting the values of u, v, x, y and the interpolation calculation control parameter of Expression (2) into Expression (4), the pixel value of the interpolation pixel is P (u, v) = 134 (decimal point or less). First
Rounding off). As a result, the pixel value of the interpolation pixel is obtained. Then, the interpolation pixel position is moved to extract the pixel values of the peripheral pixels and the interpolation processing is continued until the processing is completed for all the input original image data. When the interpolation pixel position is the same as the position of the existing pixel, the interpolation calculation is not performed and the pixel value of the original pixel is used as the pixel value of the interpolation pixel.

In the above description, the interpolation processing is performed by using four pixels as the peripheral pixels, but four or more peripheral pixels may be used. It is also possible to determine the gradation difference using four or more peripheral pixels, distinguish the edge portion, and finally perform interpolation processing using four pixels according to Expression (4). It is also possible to perform interpolation processing using four or more peripheral pixels. In this case, it is necessary to change the interpolation calculation control parameter according to the number of pixels used.

The interpolation processing method of the present invention will be described below. FIG. 9 is a flowchart showing an interpolation processing method in the interpolation processing unit 17. In step S1-1, the interpolation processing unit 17 starts the interpolation processing. In step S1-2, the original image data is input from the spatial filter processing unit 16 to the interpolation processing unit 17. The input may be every pixel, every line, or every plane. The next process is started at the time when the pixel value of a predetermined peripheral pixel for the interpolation pixel position for each plane is input. Until then, the original image data is input at any time. In step S1-3, the peripheral pixel value extraction means 51 extracts the pixel values of the peripheral pixels. Step S1-
At 4, it is determined whether the pixel value extraction is completed. When it is determined that the extraction is completed, the process proceeds to step S1-5, and when it is determined that the extraction is not completed, the step S1-
Return to 2.

In step S1-5, all combinations for selecting two pixels from the peripheral pixels included in the predetermined peripheral pixel area are obtained. In step S1-6, the gradation difference is calculated as an absolute value for each combination, and the calculated gradation difference is compared with a predetermined threshold value to determine whether the gradation difference is large or small. In step S1-7, based on the result of the determination,
Grouping is performed based on the gradation difference by grouping adjacent ones in a combination having a small gradation difference.
In step S1-8, pattern matching between the combination pattern and the predetermined table data is performed. In step S1-9, the interpolation calculation control parameter set in advance according to the collated pattern is set in the interpolation calculation formula. In step S1-10, interpolation calculation processing is performed.
In step S1-11, the interpolated image data is output. In step S1-12, it is determined whether the processing has been completed for all the input original image data. If it is determined that the processing of the image data has not been completed, the process returns to step S1-3, the interpolation pixel position is moved, the pixel values of the peripheral pixels are extracted, and the interpolation processing is continued. When it is determined that the processing is completed, step S1
Go to -13. In step S1-13, the interpolation process ends. As described above, high image quality can be maintained by preserving the sharpness for the edge portion having a large gradation difference and smoothly interpolating the flat portion having a small gradation difference.

It is also possible to use the interpolation processing method of the present invention for the brightness value and to use a combination of interpolation methods such as NN or BL for the pixel value representing the color component.

[Embodiment 2] The above-described interpolation processing is not limited because the input image can be realized in either a color image or a monochrome image, but it is particularly effective for a monochrome image. . Even in the interpolation processing of a color image, it is possible to interpolate the input image data without performing color conversion into the LCC space, using RGB signals as they are, and obtaining a gradation difference for each component (plane), but representing a color component. Even better results can be obtained by using the color difference which is the gradation difference of the pixel value (CC). Next, a second embodiment will be described in which the interpolation process is performed using not only the gradation difference of the luminance value (L) but also the color difference.

FIG. 10 is a block diagram showing the structure of the interpolation processing section 17 of the image processing apparatus 1 of the present invention. The configuration of the image processing device 1 is the same as above. The interpolation processing unit 17 includes a peripheral pixel value extraction unit 71 and a pixel selection unit 7.
2. Luminance value gradation difference determination means 73a and color difference determination means 73
b, grouping means 74, pattern matching means 75, interpolation calculation control parameter setting means 76 and interpolation calculation processing means 77.

The peripheral pixel value extraction means 71 extracts a pixel value for each peripheral pixel at the interpolation pixel position from the image data filtered by the spatial filter processing section 16.
The extraction method is the same as above. In addition to the luminance value (L), the pixel value is also extracted for the pixel value (CC) representing the color component. The pixel selection means 72 selects all the combinations that select two pixels from the peripheral pixels included in the predetermined peripheral pixel area. The selection method is the same as described above.

The brightness value gradation difference determination means 73a obtains a gradation difference for the brightness value (L) for each selected combination, compares it with a predetermined threshold value, and compares the brightness value (L) gradation difference. To judge. Then, the color difference determining unit 73b determines a color difference which is a gradation difference of pixel values (CC) representing color components for each selected combination, and compares the color difference with a predetermined threshold value to determine the color difference. As a result, even for the color edge portion visually recognized as an edge by human eyes, it is possible to precisely distinguish which is the edge portion and which is the flat portion in the peripheral pixel area.

The grouping means 74 divides the combinations into groups based on the result of the gradation difference and the color difference of the brightness values determined in the peripheral pixel area, and performs grouping. The pattern matching unit 75 matches the grouped combination patterns with the patterns set in advance as table data. The matching method is the same as above. As a result, the interpolation calculation control parameter for obtaining the pixel value of the interpolation pixel can be uniquely determined, and the optimum interpolation processing can be realized for the edge portion or the flat portion that is finely distinguished. Further, since the table data is prepared in advance, the arithmetic processing can be simplified and the cost of the interpolation processing unit 17 can be suppressed. The interpolation calculation control parameter setting means 76 sets an interpolation calculation control parameter predetermined for each table data in the interpolation calculation formula based on the collation result. Interpolation calculation processing means 77
Performs interpolation calculation processing using the interpolation calculation control parameter set by the interpolation calculation control parameter setting means 76. Finally, the output image data after the interpolation processing is output to the output gradation correction unit 110.

As described above, even when the color difference is used, the configuration of the interpolation processing unit 17 is basically the same as that of the first embodiment, but the difference is that the brightness value is used as the image data for the determination of the gradation difference. Pixel value (CC) representing not only (L) but also color components
Since it also uses, the color difference determining unit 73b is included. Accordingly, by separately determining the color difference between the two Cs, it is possible to perform the interpolation processing of the color edge portion, which is impossible by the interpolation processing of only the gradation difference of the luminance value (L). Since the interpolation calculation is performed for each LCC data,
It is necessary to separately provide a threshold value for determining the color difference. Further, it is necessary to perform the color difference determination after performing the gradation difference determination of the brightness value, and perform the grouping based on the determination result.

Next, the interpolation processing method of the present invention will be described below. FIG. 11 is a flowchart showing an interpolation processing method in the interpolation processing unit 17. In step S3-1, interpolation processing is started, and in step S3-2, the original image data is input from the spatial filter processing unit 16 to the interpolation processing unit 17. The input may be every pixel, every line, or every plane. When the pixel value of a predetermined peripheral pixel for the interpolation pixel position for each plane is input, the next process is started. Until then, the original image data is input at any time. In step S3-3, the peripheral pixel value extraction means 71 extracts the pixel values of the peripheral pixels. In addition to the luminance value (L), the pixel value is also extracted for the pixel value (CC) representing the color component. Step S3-4
Then, it is determined whether or not the pixel value extraction is completed. When it is determined that the extraction is completed, the process proceeds to step S3-5, and when it is determined that the extraction is not completed, the step S3- is performed.
Return to 2.

In step S3-5, all combinations for selecting two pixels from the peripheral pixels included in the predetermined peripheral pixel area are obtained. In step S3-6, the gradation difference and the color difference of the brightness value are calculated as absolute values for each combination, and the calculated respective values are compared with a predetermined threshold value to determine whether the gradation difference is large or small. As a result, even for the color edge portion visually recognized as an edge by human eyes, it is possible to precisely distinguish which is the edge portion and which is the flat portion in the peripheral pixel area. In step S3-7, based on the result of the determination, the adjacent ones are grouped in a combination in which the gradation difference of the brightness values is small, thereby performing the grouping based on the gradation difference. In addition, color differences are also grouped. In step S3-8, pattern matching between the combination pattern and predetermined table data is performed.

In step S3-9, the interpolation calculation control parameters set in advance in the collated pattern are set in the interpolation calculation formula. In step S3-10, interpolation calculation processing is performed, and in step S3-11, the interpolated image data is output. In step S3-12, it is determined whether the processing has been completed for all the input original image data. If it is determined that the processing of the image data is not completed, the process returns to step S3-3, the interpolation pixel position is continuously moved, the pixel values of the peripheral pixels are extracted, and the processing is continued. If you determine that the process is complete,
It proceeds to step S3-13. In step S3-13,
The interpolation process ends.

As described above, even when the color difference is used, the processing flow of the interpolation processing method is basically the same as that of the first embodiment, but the difference is that the luminance difference is used as the image data for the determination of the gradation difference. Not only the value (L) but also the pixel value (C
C) is also used. With this, by separately determining the color difference between the two Cs, it is possible to interpolate the color edge portion, which is impossible by the interpolation process of only the gradation difference of the luminance value (L). Since the interpolation calculation is performed for each LCC data, it is necessary to separately provide a threshold value for determining the color difference. Further, it is necessary to perform the color difference determination after performing the gradation difference determination of the brightness value, and perform the grouping based on the determination result.

Next, an embodiment using another arithmetic processing method of the present invention will be described. The configuration of the image processing device 1 is the same as above.

[Third Embodiment] A third embodiment of the present invention will be described below. FIG. 12 is a block diagram showing a configuration of the interpolation processing unit 17 included in the image processing apparatus 1 of the present invention. The configuration of the image processing device 1 is the same as above. The interpolation processing unit 17 includes peripheral pixel value extraction means 61, provisional pixel value setting means 62, distance calculation means 63, gradation difference calculation means 64, antecedent part membership value calculation means 65, consequent part membership value calculation means. 66 and interpolation calculation processing means 67.

The peripheral pixel value extraction means 61 extracts a pixel value for each peripheral pixel at the interpolation pixel position from the image data filtered by the spatial filter processing section 16.
The extraction method is similar to that of the first embodiment. For example, when four pixels are used as peripheral pixels,
As shown in (a), a pixel P1 located at the upper left of the interpolation pixel position P is extracted, and P1, P2, and P shown by black circles surrounding the interpolation pixel position P in the pixel region of 8 neighborhoods surrounding P1.
The four peripheral pixels of 3 and P4 are extracted, and the respective pixel values are extracted. As a result, the pixel values of the peripheral pixels can be uniformly extracted.

The provisional pixel value setting means 62 obtains the average value of the pixel values of the peripheral pixels at the peripheral pixel positions and sets it as the provisional pixel value of the interpolation pixel. Specifically, the average value of the pixel values of the peripheral pixels surrounding the interpolated pixel position P is obtained and set as the provisional pixel value p. As shown in the following formula, the extracted P1, P2, P
The sum of the pixel values p1 to p4 of the four peripheral pixels of 3 and P4 is calculated and divided by 4 to calculate the average value, which is set as the provisional pixel value p. p = (p1 + p2 + p3 + p4) / 4 (5) Note that the provisional pixel value of the interpolation pixel is a temporary pixel value of the interpolation pixel.

The distance calculating means 63 calculates the distance between the interpolated pixel position and each peripheral pixel. Specifically, the Euclidean distance will be obtained. For example, when the value d1 of the distance between the interpolated pixel position P and the peripheral pixel P1 in FIG. 3A is obtained, it is obtained by the following formula. Here, the respective coordinates are P (x, y), P1 (x1, y1), P2.
(X2, y2), P3 (x3, y3) and P4 (x
4, y4). d1 = √ ((x−x1) ² + (y−y1) ² ) ... (6)

The distance values d2, d3 and d4 between the interpolated pixel position P and P2, P3 and P4 are similarly obtained. This makes it possible to easily determine which of the peripheral pixels is closest to the interpolation pixel position P. Further, since the calculation method is not complicated, the configuration of the device that realizes the interpolation processing can be simplified, and the interpolation processing unit 1
The cost of 7 can be suppressed.

The gradation difference calculating means 64 determines the gradation difference between the provisional pixel value p and the pixel value of each peripheral pixel at any time. To calculate the gradation difference, the provisional pixel value p and the pixel values p1 to p4 of each peripheral pixel are calculated.
Find the absolute value of the difference between and. For example, the gradation difference value s1 between the provisional pixel value p and the pixel value p1 of the peripheral pixel P1 in FIG. 3A is obtained by the following formula. s1 = | p1-p | (7)

The gradation difference values s2, s3, and s4 between the provisional pixel value p and the pixel values of P2, P3, and P4 are similarly obtained. This makes it possible to determine which of the peripheral pixels is closest to the provisional pixel value p. Further, since the calculation method is not complicated, the configuration of the device that realizes the interpolation processing can be simplified, and the interpolation processing unit 1
The cost of 7 can be suppressed. Then, the distance values d1 to d4 and the gradation difference values s1 to s4 obtained above are respectively substituted into the non-linear function to calculate the value for obtaining the value of the interpolation pixel. In the present invention, the membership function used in fuzzy theory is used as the non-linear function.

The concept of fuzzy sets is used in fuzzy theory. A fuzzy set is a set concept for quantitatively expressing the ambiguity in the meaning of words and the definition of concepts. In the conventional concept of a set, whether a certain element in the meaning of the word belongs or does not belong is defined by a binary value of 1 or 0. On the other hand, in the concept of fuzzy sets, it is defined whether a certain element belongs or does not belong to the word by expanding the degree as a continuous value from 0.0 to 1.0. A value indicating this degree is called a membership value. The concept of fuzzy sets will be described with reference to FIGS. 13 and 14. FIG. 13 is a diagram showing an example of a conventional set, and FIG. 14 is a diagram showing an example of a fuzzy set. For example, considering the words "cold", "comfortable", and "hot" with respect to the temperature, in the conventional set, for example, as shown in FIG. 13, it is not comfortable if it deviates from the temperature in the range of "comfortable" even a little. I will end up. On the other hand, in the fuzzy set, as shown in FIG. 14, the degree to which each element belongs to the concept can be expressed using a membership function that takes a continuous value between 0.0 and 1.0. . In the interpolation calculation process, by using the membership function, the proximity of the interpolated pixel position and each peripheral pixel and the smoothness of the gradation between the provisional pixel value and the pixel value of the peripheral pixel are digitized as a degree, It is possible to quantitatively express the ambiguity in the definition of the concept of words such as "" and "smoothness". Inference using a fuzzy set is called fuzzy inference. Fuzzy inference is a method that allows a wide range of relationships between preconditions and conclusions by including fuzzy ideas in the inference mechanism. Therefore, by using this fuzzy inference, the goodness of fit of the fuzzy rule is calculated from each membership value based on the fuzzy set represented by two membership functions related to Euclidean distance and gradation difference, and the degree of goodness of fit It is possible to calculate a pixel value of an appropriate interpolation pixel according to In the present invention, as a fuzzy rule, a rule "if the Euclidean distance between P1 and P is short and the gradation difference between p1 and p is small, the interpolated pixel value becomes a value close to p1". . The "if-if" part of the fuzzy rule is called the antecedent part, and the part that is the conclusion is called the consequent part. Membership functions used in the interpolation calculation process of the present invention are shown in FIGS. FIG. 15 is a diagram showing a first antecedent part membership function regarding the degree of closeness of the Euclidean distance, and FIG. 16 is a diagram showing a second antecedent part membership function regarding the degree of small gradation difference. FIG. 17 is a diagram showing the consequent part membership function regarding the degree of relation of the peripheral pixel with respect to the pixel value of the interpolation pixel.

The antecedent part membership value calculation means 65 is
Distance values d1 to d calculated by the distance calculation means 63
4 as an input value, the interpolation pixel position P and each of the peripheral pixels P1 to P
The first antecedent part membership value calculation means for outputting the degree of closeness to the distance 4 as a numerical value, and the calculated gradation difference values s1 to s4 as input values, and the provisional pixel value p and each periphery. The second antecedent part membership value calculating means calculates numerically the degree of the gradation difference from the pixel values p1 to p4 of the pixels. The respective values are input to the first and second antecedent part membership functions, and the first and second antecedent part membership values are calculated. For the calculation, a calculation based on a predetermined membership function formula may be performed, or table data that is obtained in advance and stored in the memory may be referred to.

The consequent part membership value calculating means 66 is
The logical sum of the first and second antecedent part membership values calculated by the antecedent part membership value calculation means 65 is used as the input value of the consequent part membership function to calculate the consequent part membership value. . The interpolation calculation processing means 67
The pixel value of the interpolation pixel is calculated by substituting the output value calculated by the consequent part membership value calculation means 66 into the following equation. P (x, y) = (p1 × α + p2 × β + p3 × γ + p4 × δ) / (α + β + γ + δ) (8)

As a result, the vague sense of distance and the smoothness of the vague gradation can be evaluated numerically and evaluated, and the pixel of the appropriate interpolation pixel can be obtained by obtaining the degree of conformity with a predetermined fuzzy rule from these values. The value can be calculated. Finally, the output image data after the interpolation processing is output to the output gradation correction unit 110. As described above, high image quality can be maintained by preserving the sharpness for the edge portion having a large gradation difference and smoothly interpolating the flat portion having a small gradation difference.

Next, the above-mentioned interpolation processing will be concretely described using the interpolation pixel position P of FIG. First, the pixel values of peripheral pixels are extracted. The extraction method is the same as that of the first embodiment, and P1 and P indicated by black circles surrounding the interpolation pixel position P in FIG.
Pixel values of four pixels of 2, P3 and P4 are extracted.
FIG. 18 is a diagram showing a part of FIG. 4. As shown in the figure, the brightness values are 140 for P1, 132 for P2, 201 for P3, and P.
4 is 151. Moreover, each coordinate is P1 (30, 3
0), P2 (31, 30), P3 (30, 31), P4
(31, 31), and the coordinates of the interpolation pixel are P (30.
8, 30.3).

Next, based on this extraction result, the interpolation pixel position P
The provisional pixel value p of the interpolation pixel for is set. Formula (5)
Therefore, the provisional pixel value p is p = (140 + 132 + 201 + 151) / 4 = 156. When p is calculated, the first decimal place is rounded off.

Next, when the Euclidean distances d1 to d4 are calculated from the equation (6) in the order of P1 to P4, d1 = √ ((30.8-30) ² + (30.3-3
0) ² ) = 0.85, d2 = 0.36, d3 = 1.06, d4 = 0.73.

Then, when the gradation differences s1 to s4 between the provisional pixel value p and the pixel values p1 to p4 of the peripheral pixels are calculated from the equation (7), s1 = │140-156│ = 16, s2 = 24, s3 = 45 and s4 = 5. When obtaining the values of d1 to d4, the third place below the decimal point is rounded off.

Next, the antecedent part membership value is calculated. If a function expression can be defined, the membership function can be defined in advance and the membership value can be obtained by substituting each input value into the expression. If it is difficult to define it as a function, the membership value can be obtained by holding the data in advance as table data and referring to the table data corresponding to the input value. In the present invention, the following expression is used as the first antecedent part membership function f1: f1 (d) = 1.0 / (1.0 + exp (16.0 × (d−0.4))) (9 ) Is defined as the second antecedent part membership function f2, and the following expression f2 (s) = 1.0 / (1.0 + exp (0.5 × (s-20.0))) (10) is defined. Then, the membership value is calculated by substituting the input values, d1 to d4, and s1 to s4 into these expressions. Substituting each input value into f1 and f2, the membership values are: f1 (d1) = 0.00, f1 (d2) = 0.65, f1 (d3) = 0.00, f1 (d4) = 0.01, f2 (s1) = 0.88, f2 (s2) = 0.12, f2 (s3) = 0.00, f2 (s4) = 1.00. When calculating the values of f1 (d1) to f2 (s4), the third place after the decimal point is rounded off. Next, the logical sum of the first and second membership values is used as an input value and input to the consequent part membership function to obtain the consequent part membership value. Input values o1 to o4 to the consequent part membership function obtained by the logical sum are: o1 = 0.88, o2 = 0.65, o3 = 0.00, o
4 = 1.00. Substituting into the following expression f3 (o) = 1.0 / (1.0 + exp (-10.0 * (o-0.5))) (11) as the consequent part membership function f3, each peripheral pixel The degree of relation (α, β, γ, δ) with respect to P1 to P4 with respect to the interpolation pixel is obtained as follows. α = f3 (o1) = 0.98, β = 0.82, γ = 0.01, δ = 0.99

When obtaining the values of o1 to o4 and α to δ, the third place after the decimal point is rounded off. Finally,
Interpolation calculation processing is performed. Substituting the pixel values p1 to p4 and the values of α to δ of each peripheral pixel into the equation (8), P (x, y) = (140 × 0.98 + 132 × 0.82
+201 x 0.01 + 151 x 0.99) / (0.98
+ 0.82 + 0.01 + 0.99) = 142 (rounded to one decimal place). by this,
The pixel value of the interpolated pixel has been obtained. Then, the interpolation pixel position is moved to extract the pixel values of the peripheral pixels and the interpolation processing is continued until the processing is completed for all the input original image data. When the interpolation pixel position is the same as the position of the existing pixel, the interpolation calculation is not performed and the pixel value of the original pixel is used as the pixel value of the interpolation pixel.

In the above description, the interpolation calculation process is performed using four pixels as the peripheral pixels, but four or more peripheral pixels may be used. In that case, the fitness of the distance and the fitness of the gradation difference are obtained using four or more peripheral pixels, and finally, an equation obtained by expanding the equation (8), specifically, according to the number of pixels used. Values of the surrounding pixels (p1, p2, p
3, p4, ...) And the degree of relation to the interpolated pixel (α,
It is also possible to use an equation in which the number of terms of β, γ, δ, ...) Is increased.

Next, the interpolation processing method of the present invention will be described below. FIG. 19 is a flowchart showing an interpolation processing method in the interpolation processing unit 17.

In step S2-1, interpolation processing is started. In step S2-2, the original image data is input from the spatial filter processing unit 16 to the interpolation processing unit 17. The input may be every pixel, every line, or every plane. When the pixel value of a predetermined peripheral pixel for the interpolation pixel position for each plane is input, the next process is started, and the original image data is input at any time until the next process. In step S2-3, the peripheral pixel value extraction means 61 extracts the pixel values of the peripheral pixels. In step S2-4, it is determined whether the extraction of the pixel value is completed. When it is determined that the extraction is completed, the process proceeds to step S2-5, and when it is determined that the extraction is not completed, the process returns to step S2-2.

In step S2-5, the provisional pixel value of the interpolation pixel for the interpolation pixel position is set. Step S2-
In 6, the distance between the interpolated pixel position and each peripheral pixel position is calculated. In step S2-7, the gradation difference between the provisional pixel value and the pixel value of each peripheral pixel is calculated. Then, the value of the distance and the value of the gradation difference obtained above are respectively substituted into the non-linear function to calculate the value for obtaining the value of the interpolated pixel. In the present invention, the membership function used in fuzzy theory is used as the non-linear function.

In step S2-8, the antecedent part membership value is calculated. The first antecedent part membership value is calculated by using the distance value obtained above as the input value of the first antecedent part membership function. As the input value of the second antecedent part membership function, the value of the tone difference obtained above is used to calculate the second antecedent part membership value. In step S2-9, the logical sum of the first and second antecedent part membership values is input to the consequent part membership function to obtain the consequent part membership value. In step S2-10, interpolation calculation processing is performed using the consequent part membership value, and the pixel value of the interpolation pixel is calculated. In step S2-11, the interpolated image data is output. In step S2-12, it is determined whether the processing has been completed for all the input original image data. When it is determined that the processing of the image data is not completed, the process returns to step S2-3, the interpolation pixel position is moved, the pixel values of the peripheral pixels are extracted, and the interpolation processing is continued. If it is determined that the processing is completed, the process proceeds to step S2-13. In step S2-13, the interpolation process ends. As described above, the sharpness can be preserved for the edge portion having a large gradation difference, while the smooth portion can be interpolated for the flat portion having a small gradation difference.

It is also possible to use the interpolation processing method of the present invention for luminance values and to use a combination of interpolation methods such as NN or BL for pixel values representing color components.

[Embodiment 4] The above-mentioned interpolation processing is not limited because the input image can be realized in either a color image or a monochrome image, but it is particularly effective for a monochrome image. . Even in the interpolation processing of a color image, it is possible to interpolate the input image data without performing color conversion into the LCC space, using RGB signals as they are, and obtaining a gradation difference for each component (plane), but representing a color component. A better result can be obtained by the method that also uses the color difference which is the gradation difference of the pixel value (CC). Next, a fourth embodiment will be described in which the interpolation process is performed using not only the gradation difference of the luminance value (L) but also the color difference.

FIG. 20 is a block diagram showing the configuration of the interpolation processing section 17 of the image processing apparatus 1 of the present invention. The configuration of the image processing device 1 is the same as above. The interpolation processing unit 17 includes a peripheral pixel value extraction unit 81, a provisional pixel value setting unit 82, a distance calculation unit 83, a luminance value gradation difference calculation unit 84a, a color difference calculation unit 84b, an antecedent part membership value calculation unit 85, The consequent part has a membership value calculation means 86 and an interpolation calculation processing means 87.

The peripheral pixel value extraction means 81 extracts a pixel value for each peripheral pixel at the interpolation pixel position from the image data filtered by the spatial filter processing section 16.
The extraction method is the same as the above method. In addition to the luminance value (L), the pixel value is also extracted for the pixel value (CC) representing the color component. The provisional pixel value setting means 82 sets the provisional pixel value p of the interpolation pixel for the interpolation pixel position P. The distance calculation means 83 calculates the distance between the interpolated pixel position P and each peripheral pixel position. Specifically, the distance values d1 to d4 between the interpolated pixel position P and the peripheral pixels P1 to P4 are obtained from the Euclidean distance.

The brightness value gradation difference calculating means 84a calculates the gradation difference of the brightness value (L) between the provisional pixel value p and the pixel values p1 to p4 of each peripheral pixel. To calculate the gradation difference of the brightness value,
The gradation difference values s1 to s4 of the luminance values of the temporary pixel value p and the peripheral pixels P1 to P4 are obtained. The color difference calculating unit 84b calculates a color difference that is a gradation difference of a pixel value (CC) representing a color component between the provisional pixel value and the pixel value of each peripheral pixel. To calculate the color difference, the color difference value c1 between the provisional pixel value and the peripheral pixels P1 to P4 is calculated.
~ C4 is obtained. As a result, which of the peripheral pixels is closest to the provisional pixel value can be determined with respect to the pixel value representing the color component. Further, since the calculation method is not complicated, it is possible to simplify the configuration of the device that realizes the interpolation processing and reduce the cost of the interpolation processing unit 17. Then, the values obtained by substituting the distance, the gradation difference of the luminance value, and the color difference obtained as described above into the membership function are calculated. With this, regarding color difference as well, based on the fuzzy set represented by the membership function, the goodness of fit of the fuzzy rule is calculated from each membership value, and the pixel value of an appropriate interpolation pixel is calculated according to the degree of goodness of fit. can do. In the present invention, as a fuzzy rule, "if P
If the Euclidean distance between 1 and P is short, the gradation difference between the brightness values of P1 and P is small, and the color difference between P1 and P is small, the interpolated pixel value becomes a value close to P1. The rule is set. The membership function used in the interpolation calculation process of the present invention is shown in FIG. 21 in addition to FIGS. FIG. 21 is a diagram showing a third antecedent part membership function relating to the degree of the small color difference.

The antecedent part membership value calculating means 85 is
Distance values d1 to d calculated by the distance calculation means 83
4, the first antecedent part membership value calculating means for outputting the degree of the closeness of the distance between the interpolated pixel position and each peripheral pixel as a numerical value, and the value of the gradation difference of the calculated luminance value Second antecedent part membership value calculation means for calculating, as a numerical value, the degree of the gradation difference between the brightness values of the provisional pixel value and the pixel value of each peripheral pixel, using s1 to s4 as input values; The third antecedent part membership value calculating means for calculating the degree of the small color difference between the provisional pixel value and the pixel value of each peripheral pixel as a numerical value by using the color difference values c1 to c4 as input values. . The respective values are input to the first, second, and third antecedent part membership functions, and the first, second, and third antecedent part membership values are calculated. For the calculation, a calculation based on a predetermined membership function formula may be performed, or table data that is obtained in advance and stored in the memory may be referred to. The consequent part membership value calculating means 86 calculates the logical sum of the first, second and third antecedent part membership values calculated by the antecedent part membership value calculating means 85 as a consequent part membership function. Calculate the membership value of the consequent part as the input value of.
As a result, the smoothness of the gradation of the ambiguous color component can be numerically evaluated and evaluated, and the pixel value of an appropriate interpolated pixel can be calculated by calculating the degree of conformity with a predetermined fuzzy rule. The interpolation calculation processing means 87 performs interpolation calculation processing by performing calculation using the output value calculated by the consequent part membership value calculation means 86. Finally, the output image data after the interpolation processing is output to the output gradation correction unit 110.

As described above, the configuration of the interpolation processing unit 17 is basically the same as that of the third embodiment, but the difference is that
Since not only the luminance value (L) but also the pixel value (CC) representing the color component is used as the image data for the determination of the gradation difference, the color difference determination unit 84b is provided. Accordingly, by separately calculating the membership value indicating the degree of the color difference for the two Cs as well, the gradation difference of the color edge portion, which cannot be obtained by the interpolation process of only the gradation difference of the luminance value (L), is calculated. It becomes possible to output the membership value indicating the degree.
Since the interpolation calculation is performed for each LCC data, it is necessary to separately provide a membership function corresponding to the color difference. Further, in addition to the membership value of the gradation value of the brightness value and the membership value of the distance value, the logical sum is calculated including the membership value of the color difference, and the consequent membership value is calculated based on the result. There is a need.

Next, the interpolation processing method of the present invention will be described below. FIG. 22 is a flowchart showing an interpolation processing method in the interpolation processing unit 17. In step S4-1,
Start the interpolation process. In step S4-2, the original image data is input from the spatial filter processing unit 16 to the interpolation processing unit 17. The input may be every pixel, every line, or every plane. When the pixel value of a predetermined peripheral pixel for the interpolation pixel position for each plane is input, the next process is started, and the original image data is input at any time until the next process. In step S4-3, the peripheral pixel value extraction means 81 extracts the pixel values of the peripheral pixels. In step S4-4, it is determined whether the pixel value extraction is completed. When it is determined that the extraction is completed, the process proceeds to step S4-5, and when it is determined that the extraction is not completed, the process returns to step S4-2.

In step S4-5, the provisional pixel value of the interpolation pixel for the interpolation pixel position is set. Step S4-
In 6, the Euclidean distance between the interpolated pixel position and each peripheral pixel is calculated at any time. In step S4-7, the gradation difference between the luminance value (L) between the provisional pixel value and the pixel value of each peripheral pixel and the pixel value (CC) representing the color component is obtained at any time. This makes it possible to determine which of the peripheral pixels is the closest to the provisional pixel value for the color difference that is the gradation difference of the pixel value (CC) representing the color component.

In step S4-8, the first antecedent part membership value is calculated using the value of the distance obtained above as the input value of the first antecedent part membership function. Second
As the input value of the antecedent part membership function of, the value of the gradation difference of the luminance value (L) obtained above is used to calculate the second antecedent part membership value. Further, the third antecedent part membership value is calculated by using the value of the color difference obtained above as an input of the third antecedent part membership function. In step S4-9, the logical sum of the first, second, and third antecedent part membership values is input to the consequent part membership function to obtain the consequent part membership value. In step S4-10, interpolation calculation processing is performed using the output value of the consequent part membership value, and the pixel value of the interpolation pixel is calculated. In step S4-11, the interpolated image data is output. In step S4-12, it is determined whether the processing has been completed for all the input original image data. If it is determined that the processing of the image data is not completed, the process returns to step S4-3, the interpolation pixel position is continuously moved, the pixel values of the peripheral pixels are extracted, and the processing is continued. If it is determined that the processing is completed, the process proceeds to step S4-13. Step S4
At -13, the interpolation process ends.

As described above, the processing flow of the interpolation processing method is basically the same as that of the third embodiment, but the difference is that only the brightness value (L) is used as the image data for the determination of the gradation difference. Instead, the pixel value (CC) representing the color component is also used. Accordingly, by separately calculating the membership value indicating the degree of color difference for the two Cs as well, the degree of color difference of the color edge portion, which cannot be obtained by the interpolation process of only the tone difference of the luminance value (L), is calculated. It becomes possible to output the membership value shown. In addition, it is necessary to separately provide a membership function corresponding to the color difference. Further, in addition to the membership value of the gradation value of the brightness value and the membership value of the distance value, the logical sum is calculated including the membership value of the color difference, and the consequent membership value is calculated based on the result. There is a need.

Next, the image processing program of the present invention and the recording medium recording the same will be described. The present invention may be realized by a dedicated LSI (Large Scale Integration) or may be realized by a program to be executed by a computer.

The present invention is realized by an image processing program for executing the above various image processing functions. The image processing program may be directly stored in the memory of the image processing apparatus 1 or the image forming apparatus 2, or may be stored in a recording medium that can be read by being inserted into a fixed disk device or the like. The directly stored program may be a process itself executed by a general-purpose computer such as a personal computer or a workstation, or a program loaded by using a communication network such as the Internet. . Further, it may be a program generated after a processing device or the like performs processing such as decompressing data-compressed data based on the loaded program. This allows the computer to execute the interpolation processing of the present invention.

A computer system can be constructed according to the present invention. The computer system displays an image input device 21 such as a flatbed scanner, a film scanner or a digital camera, a computer on which various processes such as the image processing are performed by loading a predetermined program, and a processing result of the computer. It is composed of an image display device such as a CRT (Cathode Ray Tube) display and a liquid crystal display, and an image output device 22 such as a printer which outputs the processing result onto paper. Further, a modem or the like is provided as a communication means for connecting to a server or the like via a network.

Further, the present invention may be realized by a computer-readable recording medium recording an image processing program for causing a computer to execute various image processing functions as described above. As a result, the image processing program can be provided as a recording medium in a portable manner.

The recording medium is a memory such as a ROM (Read Only Memo) for processing by a computer.
ry) itself. Further, the program medium may be a program medium that can be read by providing a program reading device as an external storage device and inserting a recording medium therein. In any case, the stored program may be configured to be accessed and executed by the microprocessor, and in any case, the program is read and the read program is downloaded to the program storage area of the computer. It may be configured to be executed by doing so. It should be noted that this download program is stored in the main body device in advance.

Here, the recording medium is a medium which can be separated from the main body, and is a tape system such as a magnetic tape or a cassette tape, a magnetic disk such as a flexible disk or a hard disk, or a CD-ROM (Compact Disk Rea).
d Only Memory), MO (Magneto Optical), MD (Mi
ni Disks) and DVDs (Digital Versatile Disks) such as optical disks, ICs (Integrated Circu)
its) cards (including memory cards), card systems such as optical cards, mask ROM, EPROM (Erasab
le Programmable Read Only Memory), EEPROM
(Electrically Erasable Programmable Read Only Memo
ry), or a medium that fixedly carries a program including a semiconductor memory such as a flash ROM.

[0117]

As described above, according to the present invention, it is possible to use the interpolation calculation control parameters for weighting differently depending on the magnitude of the gradation difference, and the sharpness is preserved for the edge portion having the large gradation difference. On the other hand, high image quality can be maintained by smoothly interpolating the flat portion having a small gradation difference.

Further, according to the present invention, it is possible to precisely discriminate which is the edge portion and which is the flat portion in the peripheral pixel region.

Further, according to the present invention, the interpolation calculation control parameter for obtaining the pixel value of the interpolation pixel can be uniquely determined, and the optimum interpolation processing can be realized for the edge portion or the flat portion which is finely distinguished. it can. Further, since the table data is prepared in advance, the arithmetic processing can be simplified and the cost of the interpolation processing unit can be suppressed.

Further, according to the present invention, it becomes unnecessary to prepare an extra table data pattern, and the amount of memory used for storing the table data pattern can be reduced.

Further, according to the present invention, which of the peripheral pixels is closest to the interpolation pixel position can be easily determined. Moreover, since the calculation method is not complicated, the configuration of the device that implements the interpolation processing can be simplified, and the cost of the interpolation processing unit can be suppressed.

Further, according to the present invention, it is possible to determine which of the peripheral pixels is closest to the provisional pixel value. Moreover, since the calculation method is not complicated, the configuration of the device that implements the interpolation processing can be simplified, and the cost of the interpolation processing unit can be suppressed.

Further, according to the present invention, the vague sense of distance and the smoothness of the vague gradation can be numerically evaluated, and it is appropriate to obtain the degree of conformance to a predetermined fuzzy rule from these values. The pixel value of the interpolation pixel can be obtained.

Further, according to the present invention, particularly in the interpolation processing of a color image, since there is no gradation difference of the brightness value but the color is different, sharpness is also provided to a color edge portion visually recognized as an edge by human eyes. It is possible to smoothly interpolate the flat portion while saving.

Further, according to the present invention, an image obtained by preserving sharpness for an edge portion having a large gradation difference and smoothly interpolating a flat portion having a small gradation difference can be used for various purposes. it can.

Further, according to the present invention, the computer can be made to execute the interpolation processing.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an image forming apparatus 2 having an image processing apparatus 1 according to an embodiment of the present invention.

FIG. 2 is an interpolation processing unit 1 included in the image processing apparatus 1 of the present invention.
7 is a block diagram showing the configuration of No. 7.

FIG. 3A is a diagram showing an interpolated pixel position P and its peripheral pixels, and FIG. 3B is a diagram showing all combinations of selecting two pixels.

FIG. 4 is a diagram showing a part of an input original image.

FIG. 5 is a diagram showing all combinations for selecting two pixels from the peripheral pixels included in the peripheral pixel area of the interpolation pixel position P.

FIG. 6 is a diagram showing a result of performing grouping in a peripheral pixel area of an interpolation pixel position P.

FIG. 7 is a diagram showing pattern table data.

FIG. 8 is a diagram showing interpolation calculation control parameters determined corresponding to each pattern of table data.

9 is a flowchart showing an interpolation processing method in the interpolation processing unit 17. FIG.

FIG. 10 is a block diagram showing a configuration of an interpolation processing unit 17 included in the image processing apparatus 1 of the present invention.

11 is a flowchart showing an interpolation processing method in the interpolation processing unit 17. FIG.

FIG. 12 is a block diagram showing a configuration of an interpolation processing unit 17 included in the image processing apparatus 1 of the present invention.

FIG. 13 is a diagram showing an example of a conventional set.

FIG. 14 is a diagram showing an example of a fuzzy set.

FIG. 15 is a diagram showing a first antecedent part membership function relating to the degree of closeness of Euclidean distance.

FIG. 16 is a diagram showing a second antecedent part membership function relating to the degree of small gradation difference.

FIG. 17 is a diagram showing a consequent part membership function relating to the degree of relationship of peripheral pixels with respect to interpolation pixel values.

FIG. 18 is a diagram showing a part of FIG. 4;

FIG. 19 is a flowchart showing an interpolation processing method in the interpolation processing unit 17.

FIG. 20 is a block diagram showing a configuration of an interpolation processing unit 17 included in the image processing apparatus 1 of the present invention.

FIG. 21 is a diagram showing a third antecedent part membership function regarding the degree of the small color difference.

FIG. 22 is a flowchart showing an interpolation processing method in the interpolation processing unit 17.

FIG. 23 is a diagram illustrating an example of interpolation processing by NN.

FIG. 24 is a diagram illustrating an example of interpolation processing by BL.

FIG. 25 is a diagram showing an example of interpolation processing by BC.

[Explanation of symbols]

1 Image processing device 2 Image forming device 3 Digital multifunction device 11 A / D (analog / digital) converter 12 Shading correction section 13 Input tone correction section 14 Area separation processing unit 15 Color space converter 16 Spatial filter processing unit 17 Interpolation processing unit 18 color correction section 19 Black generation undercolor removal unit 21 Image input device 22 Image output device 31 External input interface 32 External output interface 51, 61, 71, 81 peripheral pixel value extraction means 52,72 pixel selection means 53 gradation difference determination means 54,74 Grouping means 55,75 pattern matching means 56,76 Interpolation calculation control parameter setting means 57, 67, 77, 87 interpolation calculation processing means 62, 82 provisional pixel value setting means 63,83 distance calculating means 64 gradation difference calculation means 65,85 Antecedent Department Membership Value Calculation Means 66,86 Consequent Department Membership Value Calculation Means 73a Luminance value gradation difference determination means 73b Color difference determination means 84a Luminance value gradation difference calculating means 84b Color difference calculation means 110 Output tone correction unit 111 gradation reproduction processing unit

Continued front page    F term (reference) 2C262 AA24 AB07 AB13 BB01 BC03                       BC05 BC13 DA09 DA11 EA04                       EA06                 5B057 AA11 BA02 CA01 CA08 CA12                       CA16 CB01 CB08 CB12 CB16                       CD06 CE11 CE17 CE18 CH07                       CH09 DA17 DB02 DB06 DB09                       DC33 DC36 DC40                 5C076 AA21 AA22 BA06 BB04                 5C077 LL04 MP01 MP08 PP02 PP10                       PP32 PP33 PP37 PP47 PP55                       PP68 PQ12 PQ18 PQ20 PQ23

Claims (15)

[Claims] 1. An interpolation processing method for obtaining a pixel value of an interpolation pixel by performing interpolation processing using pixel values of peripheral pixels, wherein a tone difference of pixel values between peripheral pixels at an interpolation pixel position is obtained, and the tone difference is obtained. And a predetermined threshold value is compared to determine a gradation difference, and when the gradation difference is larger than the threshold value, weighting is performed so that the pixel value of the interpolation pixel approaches the pixel value of the nearest neighbor pixel among the peripheral pixels, When the gradation difference is smaller than the threshold value, the pixel value of the interpolated pixel is obtained using an interpolation calculation control parameter for weighting so that the pixel value of the interpolated pixel is linearly interpolated by the pixel value of the peripheral pixel. Interpolation processing method. 2. The gradation difference is determined for each of the combinations by determining all combinations that select two pixels from the peripheral pixels included in a predetermined peripheral pixel area. 1. The interpolation processing method described in 1. 3. The combination is grouped in a peripheral pixel area based on the result of the determination, and the grouped pattern of the combination is collated with a pattern set in advance as table data. The interpolation processing method according to claim 2. 4. If the pattern of the combination does not match the pattern of the table data, the pattern of the combination is rotated by 90 degrees, and the interpolation calculation control parameter is determined according to the number of rotations until the pattern of the combination is matched. 4. The interpolation processing method according to claim 3, wherein the setting order of is changed. 5. An interpolating method for interpolating using pixel values of peripheral pixels to obtain pixel values of interpolating pixels, wherein a provisional pixel value of interpolating pixels is obtained by obtaining an average value of pixel values of peripheral pixels at interpolating pixel positions. And the gradation difference between the interpolated pixel position and the peripheral pixel and the gradation value between the provisional pixel value and the pixel value of the peripheral pixel, respectively, with respect to the provisional pixel value, the distance and the gradation difference. And a pixel value of the interpolated pixel is obtained by using a non-linear function. 6. The interpolation processing method according to claim 5, wherein the distance is a Euclidean distance between the interpolation pixel position and the peripheral pixel. 7. The interpolation processing method according to claim 5, wherein the gradation difference is an absolute value of a difference between the provisional pixel value and the pixel values of the peripheral pixels. 8. The interpolation processing method according to claim 5, wherein a membership function is used as the nonlinear function. 9. The interpolation processing method according to claim 1, wherein the gradation difference is obtained using a pixel value representing a luminance value and a color component. 10. An image processing apparatus comprising an interpolation processing section for performing interpolation processing using pixel values of peripheral pixels to obtain pixel values of interpolation pixels, wherein the interpolation processing section extracts pixel values of peripheral pixels at interpolation pixel positions. Peripheral pixel value extracting means, pixel selecting means for obtaining all combinations for selecting two pixels from the peripheral pixels included in a predetermined peripheral pixel area, gradation difference for each combination, Gradation difference determination means for comparing the tone difference with a predetermined threshold value to determine a gradation difference; grouping means for grouping the combination in the peripheral pixel area based on the result of the determination; Pattern matching means for matching the created combination pattern with a pattern set in advance as table data, and for each of the table data based on the result of the matching. And an interpolation calculation control parameter setting means for setting an interpolation calculation control parameter predetermined to the interpolation calculation formula, and an interpolation calculation processing means for performing an interpolation calculation process using the interpolation calculation control parameter. Processing equipment. 11. An image processing apparatus comprising an interpolation processing unit for obtaining a pixel value of an interpolation pixel using a pixel value of a peripheral pixel, wherein the interpolation processing unit extracts a pixel value of a peripheral pixel at an interpolation pixel position. Extraction means, provisional pixel value setting means for obtaining an average value of pixel values of peripheral pixels at the interpolation pixel position and setting it as a provisional pixel value of the interpolation pixel, distance calculation means for obtaining a distance between the interpolation pixel position and each peripheral pixel And a gradation difference calculating means for obtaining a gradation difference between the provisional pixel value and the pixel value of each peripheral pixel, and using the calculated distance value as an input value, the distance between the interpolation pixel position and each peripheral pixel First antecedent part membership value calculation means for outputting the degree of closeness as a numerical value, and the gradation value between the provisional pixel value and the pixel value of each peripheral pixel with the calculated gradation difference value as an input value Calculate the degree of smallness as a numerical value The second antecedent part membership value calculation means, the output value of the first antecedent part membership value calculation means, and the second
The pixel value of the interpolated pixel is calculated by using the output value of the consequent part membership value calculation means that uses the logical sum of the output value of the antecedent part membership value calculation means And an interpolation calculation processing unit for obtaining 12. The gradation difference is obtained using a luminance value and a pixel value representing a color component.
Or the image processing device according to item 11. 13. An image forming apparatus, comprising: the image processing apparatus according to claim 10; an image input apparatus; and an image output apparatus. 14. An image processing program for causing a computer to execute the interpolation processing method according to claim 1. Description: 15. A computer-readable recording medium in which the image processing program according to claim 14 is recorded.