JP2001319228A - Image processor and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processor and an image processing method by which the processing of image data to filter processing can surely be performed. SOLUTION: The processor is provided with a filter coefficient template 26h for holding at least two sets of filter coefficients, and a filter composition processing part 26k for extracting two sets of filter coefficients from the filter coefficient template 26h and preparing a composite filter coefficient by utilizing a zero coefficient. In this constitution, a function for preparing a filter coefficient, by which the same result as the case of performing plural times of filter processing to the image data can be provided, can be realized, thus various filter coefficients for applying various filter processing rather than the filter coefficients of a limited number can be prepared, image processing can surely be performed by the various filter coefficients, and also the filter coefficients are composted by utilizing the zero coefficient, thus the preparation of the synthetic filter coefficient can be simplified.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a digital copying machine,
The present invention relates to an image processing apparatus and an image processing system that perform filter processing in a hardware system that handles image data, such as a facsimile or a scanner, or an application program that handles image data.

[0002]

2. Description of the Related Art Conventionally, a computer system or a hardware system (image processing circuit) that drives an application program (software) to process image data read from an image input device such as a scanner is used. It is required that a suitable image be output by performing filter processing such as enhancement processing and smoothing processing. As an example of a method for performing such a filtering process, there is an image processing apparatus disclosed in Japanese Patent Laid-Open No. 149998/1994.

In this image processing apparatus, a plurality of templates are provided for an image to be processed, an evaluation value of the image to be processed is obtained, and one template is selected according to the result. I have. As a result, it is described that a filter process with an optimal frequency characteristic can be performed for each portion.

However, in the above image processing apparatus,
In order to perform appropriate filter processing on various images, it is necessary to prepare various filter coefficients. In an actual computer system or hardware system, only a limited number of filter coefficients can be implemented due to the limitation of the memory capacity and the circuit scale, so various images can be implemented using the limited number of implemented filter coefficients. A system for processing is required.

As one of the methods, there is a method of performing a filtering process on image data a plurality of times, but there is a problem that the filtering process takes a long time. As another method, there is a method in which a function of creating a synthesis filter is incorporated in a filter processing system, thereby performing various filter processes from a limited number of filter coefficients or outputting various filter coefficients. This method is performed by the following steps.

(1) Determine the maximum size of a filter used in the system.

[0007] (2) Two filters using the coefficient of the maximum size filter as a character constant and two filters using image data (hereinafter, a character constant image) representing pixel values as a character constant are converted into image data. Calculate the processing when applied.

(3) By organizing the coefficients, the coefficients of the synthesis filter when the two filters are synthesized are obtained (the coefficients of the synthesis filter are character constant expressions).

(4) The character constant expression of the coefficient of the synthesis filter is incorporated into the system.

Steps (2) and (3) when a filter processing system for synthesizing two filters having a size of 3 × 3 to synthesize a filter having a size of 5 × 5 will be described.

FIG. 18 is an explanatory diagram showing a process of obtaining a character constant expression for calculating a synthesis filter using a character constant image and a filter using coefficients as character constants.

First, a product-sum operation is performed on a character constant image with a first filter whose coefficients are represented by character constants, and coefficients g 'to s' of the intermediate character constant image are obtained. The result is the following character constant expression:

[0013] g '= aA + bB + cC + fD + gE + hF + kG + lH + mI h' = bA + cB + dC + gD + hE + iF + lG + mH + nI i '= cA + dB + eC + hD + iE + jF + mG + nH + oI l' = fA + gB + hC + kD + lE + mF + pG + qH + rI m '= gA + hB + iC + lD + mE + nF + qG + rH + sI n' = hA + iB + jC + mD + nE + oF + rG + sH + tI q '= kA + lB + mC + pD + qE + rF + uG + vH + wI r' = lA + mB + nC + qD + rE + sF + vG + wH + xI s' = mA + nB + oC + rD + sE + tF + wG + xH + yI ... formula (1) In addition, this intermediate character constant The image is subjected to a product-sum operation with a second filter whose coefficient is represented by a character constant to obtain a character constant expression m ″ after processing. Serial becomes a character constant expression.

M ″ = g′J + h′K + i′L + 1′M + m′N + n′O + q′P + r′Q + s′R Expression (2) By substituting Expression (1) into Expression (2), each character constant image The following equation is obtained by rearranging the coefficients of the character constants (a to y).

Coefficient a: AJ Coefficient b: BJ + AK Coefficient c: CJ + BK + AL Coefficient d: CK + BL Coefficient e: CL f Coefficient DJ: AM Coefficient g: Coefficient EJ + DK + BM + AN Coefficient h: FJ + EK + DL + CM + BN + AO i + CN + B + AO factor: coefficient of FL + CO k: GJ + DM + coefficient of AP l: HJ + GK + EM + DN + BP + coefficient of AQ m: IJ + HK + GL + FM + EN + DO + CP + BQ + coefficient of AR n: coefficient of IK + HL + FN + EO + CQ + BR o: coefficient of IL + FO + CR p: coefficient of GM + DP q: coefficient of HM + GN + EP + DQ r: coefficient of IM + HN + GO + FP + EQ + DR s: of iN + HO + FQ + ER t Coefficient: IO + FR Coefficient of u: Coefficient of GPv: Coefficient of HP + GQw: I + Coefficient of HQ + GR x: coefficient of IQ + HR y: will be incorporated into the system of IR ... Equation (3) the image processing apparatus equation (3) above to incorporate the synthetic function of the filter system in a conventional manner.

[0016]

However, in the above-mentioned conventional method, 1) the size of the filter to be used is limited, and 2) the steps (2) and (3) for obtaining the coefficients of the synthesis filter are performed manually. In many cases, the character constant expression of the coefficient of the synthesis filter to be obtained may be excessive due to a calculation error or the like. 3) The coefficient of the synthesis filter obtained in steps (2) and (3) is a character constant expression. Since the character constant expression becomes longer as the filter size becomes larger, there is a possibility that an error may occur when the filter is incorporated into the filter processing system, and there is a problem that it takes time to create a synthesis filter.

[0017]

SUMMARY OF THE INVENTION It is an object of the present invention to provide an image processing apparatus and an image processing system capable of easily outputting various filter coefficients for performing more various filter processes with a limited number of filter coefficients. It is to provide.

In order to solve the above-mentioned problems, an image processing apparatus according to the present invention extracts a filter coefficient template holding at least two or more sets of filter coefficients and two sets of filter coefficients from the filter coefficient templates. A filter synthesis processing unit that creates a synthesis filter coefficient by synthesizing, and when synthesizing each filter coefficient, the filter synthesis processing unit has a size larger than each size of each of the filter coefficients. It is characterized in that a zero coefficient consisting of zero is used.

According to the above-described configuration, the filter coefficient template and the filter synthesis processing section can generate a wider variety of synthesis filter coefficients using a limited number of filter coefficients. It can handle various image processing.

Further, in the above configuration, by using the zero coefficient, the synthesis of the synthesis filter coefficients from the respective filter coefficients can be simplified, the synthesis can be easily realized by the filter synthesis processing unit, and the image processing can be performed. It can be simplified.

As a result, in the above configuration, by providing the filter coefficient template and the filter synthesis processing section, the image processing can be reliably performed and the synthesis of the synthesis filter coefficients can be simplified.

In the above image processing apparatus, the filter synthesizing processing unit further extracts three or more sets of filter coefficients from the filter coefficient template, and uses the two sets of filter coefficients from the three or more sets of filter coefficients to obtain a first synthesized filter coefficient. And a first synthesis processing unit that generates a second synthesis filter coefficient from the first synthesis filter coefficient and the remaining filter coefficients.

According to the above configuration, the filter synthesis processing section has a function of creating a synthesis filter coefficient using any three or more sets of filter coefficient templates. It is possible to output more various synthesis filter coefficients for more various filtering processes.

In the above image processing apparatus, a user input unit such as a touch panel for inputting an instruction from a user and a filter synthesizing processing unit are provided with an instruction from the user input unit.
A first selection unit that selects each set of filter coefficients synthesized by the filter synthesis processing unit.

According to the above configuration, each set of filter coefficients to be synthesized by the first selection unit is selected in accordance with an instruction from the user, so that various filter coefficients for performing various filter processes are provided by the user. Can be output according to the instruction.

The image processing apparatus further includes an image input unit for inputting image data and a feature amount analyzing unit for analyzing the feature amount of the image in the image data by a feature extraction method such as a histogram and a fast Fourier transform. And a synthesis filter buffer that holds synthesis filter coefficients, and the filter synthesis processing unit has a first specification unit that specifies a set of filter coefficients to be synthesized based on a result of the feature amount analysis of the image. May be.

According to the above configuration, it is possible to output a suitable synthesis filter coefficient for performing a suitable filtering process according to the feature amount of the image in the image data by the feature amount analysis unit and the first designation unit. , Image processing can be ensured. Further, in the above configuration, since the synthesis filter coefficient can be held by the synthesis filter buffer, for example, by calling a commonly used synthesis filter coefficient from the synthesis filter buffer, the creation of the synthesis filter coefficient can be speeded up, and the image processing can be performed quickly. Can be

In the above-described image processing apparatus, a user input unit such as a touch panel for inputting an instruction from a user and a set of respective filter coefficients to be synthesized by the filter synthesis processing unit are selected according to the instruction from the user input unit. A first selection unit, an image input unit for inputting image data, a feature amount analysis unit for analyzing a feature amount of the image in the image data by a feature extraction method such as a histogram and a fast Fourier transform, and a synthesis filter coefficient And a first filter for specifying a set of filter coefficients to be synthesized based on the result of the feature analysis of the image, and a filter for each filter coefficient to be synthesized. The user can switch between specifying the set by input from the user or by specifying the result of the feature analysis of the image. The instruction from the input unit may have a selection input for selecting.

According to the above configuration, by making it possible to switch the method of designating the filter coefficients to be synthesized, it is possible to output various filter coefficients for performing more various filtering processes.

In the above image processing apparatus, at least one of an average value of density, a density histogram, and a fast Fourier transform based on a part or all of the image data is used as a feature of an image to be analyzed by the feature analyzer. A second specifying unit that specifies a set of filter coefficients to be used by combining may be provided.

According to the above arrangement, the average value of the pixel densities in the image data, the density histogram, and the fast Fourier transform are characteristic amounts that are representative of the characteristics of the image based on the image data. By specifying a set of filter coefficients to be combined by the specifying unit, it is possible to specify a suitable filter coefficient.

The image processing apparatus has a change unit that changes the strength of a filter coefficient used for synthesis in accordance with the magnitude of the feature amount of the image analyzed by the feature amount analysis unit. Is also good.

According to the above configuration, since there is a change unit,
By changing the strength of the filter coefficient according to the magnitude of the feature amount, a more suitable filter coefficient can be selected for the filtering process on the image data.

In the image processing apparatus, the filter synthesizing unit excludes, from the candidates for filter synthesis, a filter coefficient corresponding to a feature amount equal to or smaller than a certain size, for the feature amount of the image analyzed by the feature amount analyzing unit. It may have two selection units.

According to the above configuration, when even filter coefficients that are candidates for filter synthesis corresponding to small feature quantities are used for filter synthesis, processing is performed using filter coefficients as filter synthesis candidates corresponding to large feature quantities. Although there is a possibility that the image data may be adversely affected, the configuration having the second selection unit can prevent the adverse effect.

The image processing apparatus has a template instruction history memory for storing a history of a combination specified as a set of filter coefficients to be synthesized, and the filter synthesis processing unit stores the filter data stored in the template instruction history memory. Create a synthesis filter using a frequently used combination of filter coefficients specified based on the synthesis history,
When the filter combination processing is additionally stored in the filter coefficient template and the same combination is performed in the next and subsequent times, a second combination processing unit using the corresponding additionally stored filter coefficient template may be provided.

According to the above configuration, by having the second synthesis processing unit, the synthesis of the frequently used synthesis filter coefficients can be omitted, so that the processing time for generating the synthesis filter coefficients can be shortened. Image processing can be accelerated.

In the above image processing apparatus, a filter coefficient output section for displaying or outputting the output of the filter synthesis result as numerical information of the filter coefficient, and a filter processing section for performing a filter process on the image data using the filter synthesis result. An output switching unit that switches between outputting and not outputting a filter synthesis result to the filter coefficient output unit and the filter processing unit; a filter coefficient output unit that outputs the filter synthesis result;
An output control unit that outputs the signal to at least one of the filter processing units may be provided.

According to the above configuration, the output control section can selectively use the filter coefficients as a result of the filter synthesis, or the filter coefficients used in the processing after the filter processing is performed on the image data. This makes it possible to improve convenience in filtering.

In the above image processing apparatus, the sizes of the first filter coefficient and the second filter coefficient for performing synthesis are (2i + 1) × (2j + 1) and (2k + 1), respectively.
× (2l + 1) (where i, j, k, and l are positive integers), the size is (2i + 2k + 1) × (2j + 2)
When generating the combined filter coefficient of (l + 1), a size (2i + 4k + 1) in which the first filter coefficient is regarded as a pixel value
A product-sum operation is performed between the image filter coefficient and the rotation filter coefficient using an image filter coefficient of × (2j + 4l + 1) and a rotation filter coefficient obtained by rotating the second filter coefficient, and the center of the result of the product-sum operation ( 2i + 2k + 1) × (2j
+ 2l + 1) may be provided as a third synthesis processing unit that uses the calculation result as a synthesis filter coefficient.

According to the above configuration, when the filter synthesis is incorporated in the filter processing unit, there is no need to perform a complicated operation using a character constant expression, and the filter processing unit is constructed without limiting the filter size to be used. It is possible to do.

In the above image processing apparatus, the first filter coefficient is multiplied by a first constant in which all coefficients of the first filter coefficient are integers before generating the image filter coefficient, and the second filter coefficient is generated before the rotation filter coefficient is generated. The second filter coefficient is multiplied by a second constant in which all of the coefficients are integers, and a product-sum operation of the image filter coefficients is performed by an integer operation on the rotation filter, and the center of the operation result is (2i + 2k + 1) × (2j + 2
A fourth synthesis processing unit may be provided which uses the result of dividing the calculation result of (l + 1) by the product of the first constant and the second constant as a synthesis filter coefficient.

According to the above configuration, the product synthesis operation performed in the process of the filter synthesis process is performed by the integer operation, so that the fourth synthesis processing unit can perform the filter synthesis process at a high speed, thereby speeding up the image processing. it can.

In the above image processing apparatus, the sizes of the first filter coefficient and the second filter coefficient for synthesizing are (2i + 1) × (2j + 1) and (2k + 1), respectively.
× (2l + 1), the size (2i + 4k + 1) × (2j + 4l) assuming the first filter coefficient as a pixel value
+1) using an image filter coefficient and a rotation filter obtained by rotating the second filter coefficient,
Size of image filter coefficient and synthesis filter coefficient [(2 m
+1) × (2n + 1)] (where i, j, k, l, m, and n are positive integers), and the second
By dividing each coefficient of the filter coefficient by the partial normalization coefficient, the size of the synthesis filter coefficient is (2m + 1) × (2n + 1), which is independent of the size of the first filter coefficient and the size of the second filter coefficient. May be provided.

According to the above configuration, the sizes of the first filter coefficient and the second filter coefficient to be synthesized in the above-described method are (2i + 1) × (2j + 1) and (2k +
1) × (2l + 1), the size is (2i + 2)
While k + 1) × (2j + 2l + 1) synthesis filter coefficients are obtained, the size of the synthesis filter coefficients after synthesis can be determined independently of the filter coefficients before synthesis.

The image processing apparatus may have a sixth synthesis processing unit that stops the synthesis processing of the filter when 0 is obtained as the partial normalization coefficient. According to the above configuration, the filter coefficients before synthesis can be reviewed by interrupting the filter processing.

In the image processing apparatus, when 0 is obtained as the partial normalization coefficient, the fourth synthesis processing unit or the fifth synthesis processing unit generates a synthesis filter coefficient, and generates the synthesis filter coefficient from the synthesis filter coefficient. A seventh combination processing unit that creates a target filter coefficient having a size to be used may be provided. According to the above configuration, the processing can be continued even when the partial normalization coefficient becomes 0, so that a delay in the filter synthesis processing can be avoided.

The image processing apparatus has two or more of the third to seventh synthesis processing units for synthesizing the filter coefficients, and has a third selection unit for selecting a filter synthesis process. You may.

According to the above configuration, each filter synthesizing means can be selectively used, and more various filter coefficients can be used, so that it is possible to cope with various images.

According to another image processing apparatus of the present invention, in order to solve the above-mentioned problem, a filter coefficient template holding at least two or more sets of filter coefficients is provided for image processing. When a filter synthesis processing unit that extracts and synthesizes a set of filter coefficients to create a synthesis filter coefficient synthesizes each filter coefficient, the size of each filter coefficient is greater than the size of each filter coefficient. A data input / output unit for communicating with an external device having at least a part of the image processing is provided, and the data input / output unit is provided through the data input / output unit. The control unit communicating with the external device is provided so that at least a part of the image processing function is processed by the external device. It is a symptom.

According to the above configuration, since the data input / output unit and the control unit are provided, it is possible to perform image processing with higher accuracy than the image processing performed inside the image processing apparatus by an external device. Processing can be performed with higher precision. In addition, a cost advantage can be obtained by sharing the external device.

According to another aspect of the present invention, there is provided an image processing system comprising: the image processing apparatus; and an external device communicably connected to the image processing apparatus. The image processing apparatus further includes at least an image input unit that inputs image data, and a feature amount analysis unit that analyzes a feature amount of the image in the image data by a feature extraction method such as a histogram and a fast Fourier transform. The external device is provided with a feature amount correlation processing unit that obtains a feature amount having a higher correlation than the plurality of feature amounts extracted by the feature amount analysis unit.

According to the above configuration, image processing with higher accuracy than image processing performed inside the image processing apparatus can be performed by an external device, so that image processing can be performed with higher precision. In addition, a cost advantage can be obtained by sharing the external device.

In addition, since a highly correlated feature amount can be used in the entire image data, image processing suitable for the entire image data can be performed.

In the above-described image processing system, the external device further includes a filter selection unit that selects a set of filter coefficients from the filter template based on the feature amount having a strong correlation, and determines that there is no optimum filter coefficient in the filter template. In this case, a new filter generation unit that newly generates a filter coefficient may be provided.

By the way, as a filter template prepared in advance, a set of filter coefficients which is considered to have a high utilization rate under a certain assumed use condition is set, but the actual use condition differs from the assumed use condition. In
A filter template prepared in advance may not be appropriate.

However, according to the above configuration, the provision of the new filter generation unit enables automatic reconfiguration of the configuration of the filter template suitable for the actual use conditions, so that the image processing can be optimized.

In the image processing system, an external device may be provided with a filter synthesizing unit for synthesizing a filter based on the selected filter coefficient.

According to the above configuration, the filter synthesis can be performed by the external device. Therefore, when the image processing in the image processing device is reduced, the filter synthesis is performed in the external device and the processing in the image processing device is performed. If there is enough capacity, the image processing apparatus can be selectively used, for example, by performing filter synthesis inside the image processing apparatus, so that the image processing can be made more efficient and faster.

In the above-mentioned image processing system, the external device is provided with a statistical processing section for summing up the processing contents of the image processing apparatus and obtaining an optimal combination of templates corresponding to the feature amounts extracted by the image processing apparatus. May be.

According to the above configuration, the statistical processing unit can determine the optimum combination of filter templates as a system including the image processing device and the external device, so that the image processing can be optimized.

[0062]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First Embodiment A first embodiment of the image processing apparatus according to the present invention will be described in detail with reference to FIGS. 1 to 11 and the following embodiments. I do.

The first embodiment of the image processing apparatus according to the present invention is applied to a hardware system such as a digital copying machine, a facsimile, a scanner, etc., which handles images or an application program which handles images (parts which handle filter processing). The spatial filter processing unit of the image processing device used in the digital copying machine shown in (1) and a filter processing system that performs synthesis processing of filter coefficients using a computer)
Is possible.

First, a basic configuration of a digital copying machine in which the above image processing apparatus is applied as a color image processing apparatus 2 as shown in FIG. 2 will be described. FIG. 2 is a block diagram showing the entire digital copying machine. The digital copying machine has a color image input device 1, a color image processing device 2, and a color image output device 3.

The color image input device 1 is, for example, a CCD
(Charge Coupled Device), and the reflected light image from the original is converted into RGB (R: red
G: green; B: blue) are read by the CCD as analog signals.

In the color image processing device 2, first, each analog signal from the color image input device 1 is converted into digital image data by an A / D (analog / digital) converter 21. Subsequently, the shading correction unit 22
The processing for removing various distortions occurring in the illumination system, the imaging system, and the imaging system of the color image input apparatus 1 is performed. afterwards,
The input tone correction unit 23 performs a process of converting each image data, which is each reflectance signal of RGB, into a signal format such as a density signal, which is easy to handle by the image processing system, while adjusting color balance.

Next, in order to realize the fidelity of color reproduction, the color correction section 24 separates the above-mentioned image data into CMY (C: cyan, M: magenta, Y: yellow) color materials. A process for removing color turbidity based on characteristics is performed. Thereafter, the black generation / under color removal unit 25 outputs the color-corrected C
Black generation for generating black (K) image data from each image data corresponding to the three colors of MY, and subtracting the image data corresponding to K obtained by black generation from each image data corresponding to the original CMY to generate a new image. A process of generating each image data corresponding to CMY is performed, and each image data corresponding to three colors of CMY is converted into each image data corresponding to four colors of CMYK.

Next, in the spatial filter processing section 26, each of the obtained image data is subjected to a spatial filter process by a digital filter process, and the spatial frequency characteristic is corrected, thereby blurring or graininess degradation of the output image. Is processed to prevent Subsequently, the output tone correction unit 27 performs an output tone correction process of converting a signal such as a density signal into a dot area ratio, which is a characteristic value of the color image output device 3, and finally performs a tone reproduction process. In the unit 28, a tone reproduction process (halftone generation) is performed in which each image data is divided for each pixel and processed so that each tone can be reproduced.

After the color correction processing by the above-described color correction section 24, the area separation processing section 29 sets a black character (in some cases, in order to enhance the reproducibility of black characters or color characters in a mixed document including characters and photographs). In the image region extracted as the color region (including the color characters), the enhancement amount of the high frequency is increased by the sharpness enhancement process in the spatial filter processing unit 26.
At the same time, the tone reproduction processing section 28 is configured to select binarization or multi-value processing on a high-resolution screen suitable for reproduction of high-frequency components.

On the other hand, for the area determined to be a halftone dot by the area separation processing section 29, the spatial filter processing section 26 performs low-pass filter processing for removing the input halftone dot component, and In the reproduction processing unit 28, binarization or multi-value processing is performed on a screen that emphasizes gradation reproducibility.

The image data subjected to each of the above-described processes is temporarily stored in storage means (not shown), read out at a predetermined timing, and input to the color image output device 3. The color image output device 3 outputs image data on a recording medium (for example, paper).
A mono-color / color image output device using an electrophotographic system or an ink-jet system can be used, but is not particularly limited.

FIG. 1 is a block diagram of one embodiment of a filter processing system in the color image processing apparatus 2. Each block of the filter processing system includes a CPU (Central Processing Unit) 26a as shown in FIG.
Is controlled by an instruction from. The filter processing system is incorporated, for example, as a part of pre-processing in the spatial filter processing unit 26 in the block diagram shown in FIG. 2, and after an appropriate filter coefficient is generated by selection or synthesis, as shown in FIG. First, the filter processing is performed on the image data read from the image input unit 26b.

Next, a method of synthesizing filter coefficients in the above-described filter processing system will be described with reference to FIG. First, the image data input from the image input section 26b is stored in the image memory 26c [arrow C in the figure.
1. Hereinafter, the data flow is indicated by a code Cn (n is an integer)]. The image input unit 26b is a color image input apparatus 1 which is an apparatus for inputting an image such as a scanner described above.
And CD-ROM (Compact Disc-Read Only Memory)
For example, a file input / output device for inputting / outputting an image data file stored in a storage device such as a storage device can be used.

A part or all of the image data stored in the image memory 26c is converted into a FF by the feature amount analysis unit 26d.
T (Fast Fourier Transformation: Fast Fourier Transform) and feature amount analysis processing of image data such as a color histogram are performed (C2), and a template coefficient unit 26e automatically designates a set of filter coefficients based on the analysis result. (C4 → C7), and are stored in the template instruction stack 26g in the reverse order of the order of synthesis via the input switching unit 26f (C8) (first designation unit, second designation unit).

At this time, the template instruction stack 26
The numerical value stored in g is the filter coefficient template 26h
Is stored. Hereinafter, the numerical value is called a template number. An instruction to store the filter coefficient in the template instruction stack 26g can also be specified from the user input unit 26i (arrows C6, C3, and C).
5 becomes effective) (first selection unit).

The user enters a user input unit 26 composed of a touch panel (including display means such as a liquid crystal display) and the like.
Using i, information such as the type of filter coefficient used for processing is input, and stored in the template instruction stack 26g in the reverse order to the order in which the set of filter coefficients is synthesized based on the input information.

The input switching unit (selection input unit) 26f
Instruction from template instruction unit 26e and user input unit 2
Which of the instructions from 6i is the template instruction stack 2
This is for switching whether to store in 6g. The set of template numbers stored in the template instruction stack 26g is stored in the template instruction history memory 26j (C11).
If there is not enough free space to store data, the oldest template instruction history is erased in order from the oldest one until the storage becomes possible, and then stored.

At the start of filter synthesis, the filter synthesis processing unit 26k outputs 2 bytes from the template instruction stack 26g.
Two template numbers are sequentially taken out (C10), and each filter coefficient corresponding to each template number is set as a first filter coefficient and a second filter coefficient, respectively.

The set first filter coefficient and second filter coefficient are synthesized by the filter synthesis processing section 26k, and the synthesized filter coefficient is output to the synthesis filter buffer 26l (C15). The filter combination processing unit 26k confirms whether or not the template number remains in the template instruction stack 26g and whether or not the number of combined filter coefficients has reached the maximum.

If the template number remains in the template instruction stack 26g and the maximum number of filter synthesis sets has not been reached, the filter synthesis processing is repeated in the following procedure until this condition is not satisfied (first synthesis). Processing section).

After setting the filter coefficient stored in the synthesis filter buffer 26l as the first filter coefficient (C15), one template number is extracted from the template instruction stack 26g.

The filter coefficient corresponding to the template number is set as the second filter coefficient, the first filter coefficient and the second filter coefficient are combined, and a new combined filter coefficient is output to the combining filter buffer 26l.

After the filter coefficient synthesizing process to be performed by the filter synthesizing unit 26k is completed, the contents of the template instruction history memory 26j are checked (C14), and the combination of the filter coefficients synthesized this time has exceeded a certain number of times in the past. If used (for example, five or more times), the synthesis filter coefficients stored in the synthesis filter buffer 26l are added to the filter coefficient template 26h (C17 · C
16) (Second synthesis processing unit).

The synthesis filter coefficient stored in the synthesis filter buffer 26l passes through the output switching unit 26m,
Filter coefficient output unit 26 for display, printer, etc.
n, output to one or both of the filter processing units 26o such as the spatial filter processing unit 26 of the color image processing device 2 (C19, C21, C22) (output control unit). Note that arrows C9, C12, C18, C20, and C23 indicate access to each unit.

The configuration described above is a system for performing a filtering process using a computer, and the filter coefficient is output from a filter coefficient output unit 26n such as a display or a printer.
And a filter processing unit 26o such as a spatial filter processing unit 26 of the color image processing apparatus 2.

When the present filter processing system is applied to the spatial filter processing unit 26 of the color image processing apparatus 2, the user input unit 26i, the input switching unit 26f,
The output switching unit 26m and the filter coefficient output unit 26n become unnecessary.

Next, information such as the type of filter coefficient used for the filter processing is input to the user input unit 26 such as a touch panel.
Each example of inputting using i will be described with reference to FIGS.

FIG. 3 shows an embodiment of a screen 26p in which the user specifies a filter coefficient to be used. On the left side of the screen 26p shown in FIG. 3, an image to be processed and various feature amount analysis results of the image can be selectively displayed by the CPU 26a and the feature amount analysis unit 26d. The user can display the screen 26p based on this information. By manipulating the right side, selection of the filter coefficient to be used can be specified. In the example of FIG. 3, a density histogram for each color component of each input image data, Red: Red, Green: Green, Blue: Blue is shown as the feature amount of the image in each image data.

FIG. 4 shows an embodiment of a screen 26p on which a user instructs a method of selecting a filter coefficient. As shown in FIG. 4, the user can switch the method of selecting the filter coefficient by selecting a manual / automatic button on the screen 26p. When the manual button is pressed, the screen 26p in FIG.
Is displayed, and a filter coefficient to be used can be specified.

When the automatic button is pressed, each type of filter coefficient shown on the right side of the screen 26p can be selected, and it becomes possible to specify the type of filter coefficient which is a candidate for the synthesis of filter coefficients. 3 to 6,
Buttons indicated by oblique lines indicate that the buttons are currently selected.

FIG. 5 shows a screen 26 for the user to select whether to output the filter synthesis result to the output destination or not.
It is an example of p. The filter synthesis result is output to the output destination selected on the screen 26p. For example, on the screen 26p, output to the filter processing unit and the touch panel is performed.

FIG. 6 shows an embodiment of a screen 26p for the user to select a filter synthesis method. When "no limit on filter size" is selected, the method described later is used. When “limit the filter size” is selected, the size specified in the column of the maximum filter size is processed as the maximum size of the combined filter coefficient. When "limit the filter size" is selected, it is possible to select a process when the partial normalization coefficient becomes zero. Here, when "interrupt the process" is selected, the synthesis process of the filter coefficient is stopped.

On the other hand, when "continue processing" is selected, the synthesis filter coefficients are synthesized by a method described later, and a synthesis filter coefficient having a filter size set as a limit is created from the synthesized synthesis filter coefficients. .

Next, a method of synthesizing filter coefficients will be described with reference to FIGS. FIG. 7 is an explanatory diagram showing a procedure of a filter combining method in a case where, for example, 3 × 3 filter coefficients are used for combining. Such a filter coefficient synthesizing method (third synthesis processing unit) is performed in the following procedure.

1. The size of the first filter coefficient is set to (2i +
1) The values of i and j when x (2j + 1) are obtained.

2. The size of the second filter coefficient is set to (2k +
1) The values of k and l when x (2l + 1) are obtained.

3. (2i + 4k + 1) × (2j + 4l +
A storage area (hereinafter, a real image filter coefficient and a real image filter coefficient) in which a coefficient of a first filter coefficient is substituted for a center of a storage area (hereinafter, referred to as a real image zero coefficient) in which all coefficient values of a real number storage area of 1) are zero. Call). In such real number image filter coefficients, zero coefficients are arranged around a matrix of the first filter coefficients.

4. A new filter coefficient (hereinafter, referred to as a rotation filter) is created by moving (rotationally converting) each coefficient of the second filter coefficient around the center of the second filter coefficient, for example, to a position rotated by 180 degrees.

5. Center of real image filter coefficients (2i
+ 2k + 1) × (2j + 2l + 1) is subjected to the product-sum operation of the rotation filter.

6. (2i + 2k) obtained by the product-sum operation
A coefficient having a size of (+1) × (2j + 2l + 1) is a combined filter coefficient obtained by combining the first filter coefficient and the second filter coefficient. FIG. 7 illustrates a process when i = j = k = 1 = 1 in the above-described filter synthesis method.

FIG. 8 shows the procedure of another filter synthesis method.
FIG. 3 is a diagram illustrating a case where, for example, 3 × 3 filter coefficients are used for synthesis. The filter synthesis method (fourth synthesis processing unit) is performed in the following procedure.

1. The size of the first filter coefficient is set to (2i +
1) The values of i and j when x (2j + 1) are obtained.

2. The size of the second filter coefficient is set to (2k +
1) The values of k and l when x (2l + 1) are obtained.

3. A constant a is obtained such that when all coefficients of the first filter coefficient are multiplied by a constant a, all a × coefficients are integers.

4. A constant b is obtained such that when all coefficients of the second filter coefficient are multiplied by a constant b, all b × coefficients are integers.

5. (2i + 4k + 1) × (2j + 4l +
A storage area in which the value obtained by multiplying each coefficient of the first filter coefficient by a constant a is substituted in the center of a storage area (hereinafter, referred to as an integer image zero coefficient) in which all values of the integer storage area of 1) are set to 0 (hereinafter, referred to as an integer image zero coefficient). , Called integer image filter coefficients).
In this integer image filter coefficient, zero coefficients are arranged around a matrix of first filter coefficients multiplied by a.

6. Each coefficient of the second filter coefficient is multiplied by a constant b, and around the center of the second filter coefficient, for example, 180
A new filter coefficient (hereinafter, referred to as an integer rotation filter) moved to the position rotated by degrees is created.

7. Center of integer image filter coefficients (2i
+ 2k + 1) × (2j + 2l + 1) is subjected to the product-sum operation of the integer rotation filter.

8. (2i + 2k) obtained by the product-sum operation
+1) × (2j + 2l + 1) is divided by ab.

9. The obtained coefficient is used as a coefficient of a combined filter coefficient obtained by combining the first filter coefficient and the second filter coefficient. FIG. 8 shows that i = j in the filter synthesis method described above.
= K = l = 1 is illustrated as an example.

FIG. 9 is a diagram showing a procedure of still another filter synthesizing method when, for example, 3 × 3 filter coefficients are used for synthesis. The filter synthesis method (fifth synthesis processing unit) is performed according to the following procedure.

1. The size of the first filter coefficient is set to (2i +
1) The values of i and j when x (2j + 1) are obtained.

[0113] 2. The size of the second filter coefficient is set to (2k +
1) The values of k and l when x (2l + 1) are obtained.

3. The size of the synthesis filter coefficient to be output as a result (hereinafter, referred to as a mask size) is (2m + 1)
The value of m and n when x (2n + 1) is obtained.

4. (2k + 2m + 1) × (2l + 2n +
A storage area in which the coefficient of the first filter coefficient is substituted (hereinafter, a real image filter coefficient) with respect to the center of a storage area (hereinafter, referred to as a real image zero coefficient) in which the values of the real number storage area in 1) are all set to 0 Call). At this time, the real image filter coefficients are arranged such that zero coefficients surround the matrix of the first filter coefficients.

5. A new filter coefficient (hereinafter, referred to as a “transformation”) that is obtained by moving each coefficient of the second filter coefficient around the center of the filter coefficient, for example, to a position (rotation conversion) rotated by 180 degrees.
(Referred to as a rotation filter).

6. Sum of pixel values of real image filter coefficients included in a (2m + 1) × (2n + 1) window opened at the position of each coefficient of the rotation filter when the centers of the rotation filter and the real image filter coefficients are aligned (hereinafter, , A partial normalization coefficient), and a synthesis filter coefficient (hereinafter, referred to as a partial normalization rotation filter) is created with a value obtained by dividing each coefficient of the rotation filter by the partial normalization coefficient as a new coefficient.

However, if 0 is obtained as the partial normalization coefficient, the filter synthesis processing is stopped as an error (sixth synthesis processing unit). Alternatively, when “limit the filter size” is selected in the filter coefficient combining method and “continue processing” is selected as the processing when the partial normalization coefficient becomes 0 (see FIG. 6), The synthesis filter coefficient (target filter coefficient) is synthesized by the method shown in FIG. 7 or FIG. 8 (seventh synthesis processing unit).

7. Center of real image filter coefficients (2m
+1) × (2n + 1) and a partially normalized rotation filter.

8. (2m + 1) obtained by the product-sum operation
A coefficient having a size of × (2n × 1) is set as a coefficient of a combined filter coefficient of the first filter coefficient and the second filter coefficient.

FIG. 9 exemplifies the processing when i = j = k = 1 / m = n = 1 in the above-described filter synthesis method. Further, the respective filter combinations may be set to be arbitrarily combined and selected (third selection unit).

As described above, each of the filter synthesizing methods described above corresponds to (2) in the prior art "method of performing various filtering processes from a limited number of filter coefficients or outputting various filter coefficients". The result of the trial and error analysis of the step (3) (refer to the section of the prior art) and the result of the trial and error made by the present inventor is that the filter synthesis processing unit 26k
However, when synthesizing each filter coefficient, a zero coefficient having a size larger than each size of each filter coefficient and including all zeros is used.

With such a configuration, the synthesis of the synthesis filter coefficients from the respective filter coefficients can be simplified, and the synthesis can be easily incorporated into the filter synthesis processing unit 26k, and the image processing can be simplified. Since it is possible to generate more various synthesis filter coefficients and perform more various filtering processes, image processing can be more reliably performed.

FIG. 10 shows a fast Fourier transform (hereinafter referred to as FF) as a feature used in each of the above-described filter synthesizing methods.
FIG. 9 is a diagram showing an example in which T is used. FIG.
The graph 0 (a) shows the result of the FFT processing performed by the feature amount analysis unit 26d, and shows a case where the feature amounts (1), (2), and (3) appear as feature amounts. FIG.
The threshold value shown in (a) is for determining which feature value is to be used, and is appropriately set according to the processing content and the like.

The graphs in FIGS. 10B and 10C show the frequency characteristics of the filter coefficients which are the filter synthesis candidates corresponding to the feature amounts (1), (2) and (3). (In this example, filter candidate 1. indicates a high-pass filter, and filter candidate 2. indicates a low-pass filter.)

The strengths (amplitudes) of the feature quantities are in the order of feature quantity (1)> feature quantity (2)> feature quantity (3), and the magnitudes of the filter synthesis candidates corresponding to the feature quantities have the same magnitude relation. However, the filter synthesis candidate corresponding to the feature value (3) is not used for synthesis because the strength of the feature value (3) is smaller than the threshold value (second selecting unit).

FIG. 11 is a diagram showing an embodiment in which a histogram is used as a feature value used in each of the above-described filter synthesizing methods. Each of the graphs in FIG. 11 is a feature amount analysis unit 26d when (a) a full-color image such as a photograph, (b) a black-and-white document centered on characters, and (c) a business document including a colored diagram or graph are input images. 5 is a graph showing a typical histogram shape generated in FIG.

As described above, since the tendency of the histogram differs depending on the type of image in the image data, filter characteristics suitable for each type of image (for example, for a full-color image, the filter coefficient for smoothing is set to be strong, and conversely, By using a filter coefficient having an emphasis filter coefficient for other images) for synthesis, it is possible to perform filter processing suitable for the type of image. When the filter coefficient is set to be strong or weak, a constant (positive real number) to be multiplied by the filter coefficient is set in a plurality of stages, and when a large constant is multiplied by the CPU (change unit) 26a, the filter coefficient is set strongly. Multiplying by a small constant will result in a weak setting.

When a histogram and an average value are used as the feature amounts used in the above-described respective filter synthesizing methods, the type of the image is determined by the above-described method, and the image data within the small range of the image in the image data is determined. RGB of average value
From the balance, it is determined whether the range is color or monochrome. As a result, by classifying in more detail what kind of image the attention range of the image data is, and using a filter coefficient having filter characteristics suitable for the image type (including color / monochrome information) for synthesis, Filtering suitable for the type of image can be performed.

[Second Embodiment] The following will describe a second embodiment of the image processing apparatus of the present invention with reference to FIGS. In the second embodiment, members having the same functions as those in the first embodiment are given the same member numbers, and descriptions thereof are omitted.

In the second embodiment, FIG. 12 and FIG.
As shown in FIG. 3, the color image processing device 2 as an image processing device is connected to an image processing server 43 as an external device via a bus line 42, and is subjected to image processing by a control unit (for example, the CPU 26a shown in FIG. 12). For example, information for image processing such as transmission / reception of image feature amounts, transmission / reception of filter template numbers, and filter coefficients can be mutually transmitted / received.

For this reason, the color image processing apparatus 2 includes:
An external device connection unit 4 as a data input / output unit is provided as an interface. Therefore, in the color image processing apparatus 2, at least a part of the image processing can be executed by the image processing server 43 as an external device by the control unit.

For connection via such a bus line 42, a wired or wireless LAN (Local Area Network) is used.
work) or a connection form in which the number of connection destinations such as a telephone public line is unspecified and large. Further, an image processing server 44 having the same function as the image processing server 43 is further connected to the bus line 42, and two or more image processing servers 43, 44,
The setting may be made so that communication with the color image processing apparatus 2 is possible. By connecting the color image processing device 2 to the image processing server 43 and the like in this way, an image processing system according to the present invention can be constructed.

Here, a case where the image processing server 43 is connected as an external device will be described. When the image processing server 43 is not used, the input image data 45 is divided into, for example, n areas in the color image processing apparatus 2 due to restrictions on hardware scale and cost, as shown in FIG. The analysis of the image feature amount for each region, for example, from the region 46 of the image data 45, the image feature amount (1),
Analysis of the image feature amount (n) from the area 47 of the image data 45;
Then, a filter template to be synthesized is selected.

Similarly, the selection of the filter template is also restricted by the hardware scale and cost, as described above.
This is performed based on simple criteria such as the peak position and the peak height.
When the processing is performed only in the color image processing apparatus 2 as described above, the selection of the filter template is performed for each area, so that when the image data is viewed as a whole, it may not always be optimal.

When the image processing server 43 is used, the feature amounts (1) to (n) obtained by the color image processing apparatus 2 are used.
Is transmitted to the image processing server 43 via the connected bus line 42, and is appropriately processed in the image processing server 43 (C24 and C25 in FIG. 12).

As shown in FIG. 15, the image processing server 43 has a feature amount correlation processing section 48, a filter selection section 49, a new filter creation section 50, a filter synthesis section 51, and a statistical processing section 52. In the image processing server 43, the feature values (1) to (n) are input to the feature value correlation processing unit 48, and a set X of feature values having a strong correlation with each other is obtained.

In the feature amount correlation processing section 48, as shown in FIG. 16, for example, a feature amount exceeding a predetermined threshold value for a feature amount of each region is subjected to frequency distribution for the entire image data. By calculating and extracting a feature amount having a large frequency in the frequency distribution, a set X of feature amounts having strong characteristics with respect to the entire image data is obtained.

Alternatively, as shown in FIG. 17, a correlation coefficient between a characteristic amount for each region and a comparison characteristic pattern prepared in advance is calculated, and the correlation coefficient is set to a predetermined threshold value. A frequency distribution is calculated for the entire image data for a feature amount exceeding α, and a feature amount having a large frequency in the frequency distribution is extracted by the threshold α, whereby a feature amount having a strong characteristic throughout the entire image data is obtained. Get the set X.

The feature amount set X is transmitted to the color image processing device 2 and used for selecting a filter coefficient template (C26 and C28 in FIG. 12). Alternatively, the set X of the feature amounts is stored in the filter selection unit 49 in the image processing server 43.
Is output to

The filter selection section 49 of the image processing server 43
In, an optimal set of filter coefficients is selected from a filter template based on the set X of feature amounts. At this time, if there is no filter coefficient required as an optimal set of filter coefficients in the filter template, the new filter creation unit 50
Generates an optimum filter coefficient.

In this process, all of the selected or created sets of filter coefficients are used as templates in the filter coefficient template 26h in the color image processing apparatus 2 as the optimum set for the obtained feature quantity set X. If it is held, the template number indicating that is held is transmitted to the color image processing device 2.

If not all the sets of the filter coefficients are stored as templates in the color image processing apparatus 2, the filter coefficients and the template numbers indicating the filter coefficients are transmitted to the color image processing apparatus 2.

The color image processing device 2 and the image processing server 4
3, data is transmitted and received between each other, and the above processing is determined by the CPU (not shown) of the image processing server 43.

Also, the obtained filter coefficient set is not transmitted to the color image processing apparatus 2, but a filter is synthesized from the filter coefficient set in the image processing server 43 (filter synthesis unit 51), and the combined filter coefficient is obtained. Color image processing device 2
May be transmitted. In this case, the combined filter coefficients are stored in the combined filter buffer 26 of the color image processing apparatus 2.
1 is stored in the filter coefficient template 26h (C27 in FIG. 12).

It should be noted that whether or not the above-described image processing server 43 is used, whether or not the template selection based on the characteristic amount X is performed by the image processing server 43,
Whether or not to obtain the synthesis filter in 3 is input in advance from the user input unit 26i of the color image processing apparatus 2.

The image processing server 43 is provided with a statistical processing section 52, and the processing contents of all the image processing apparatuses such as the color image processing apparatus 2 connected to the image processing server 43 are totaled ( For example, a combination of a feature amount and a template that is more suitable as a filter template prepared in the image processing apparatus is required. The combination of the requested filter template is
It is transmitted to each image processing device using the idle time of the processing of the image processing server 43 and each image processing device.

In the above, an example in which the image processing apparatus having the filter synthesizing function described in the first embodiment is used as the image processing apparatus has been described. However, an image forming apparatus such as a generally used digital color copying machine is used. An image processing device provided in the device may be connected.

In this case, the feature amount transmitted from the image forming apparatus is received, and a set of templates selected based on the total result of the plurality of image processing apparatuses (of the first embodiment) is received. The combined filter coefficients are configured to be transmitted.

That is, by providing the statistical processing section 52 in the image processing server 43, it is possible to obtain the optimum filter coefficient based on the characteristic amount even for the image processing apparatus provided in the image forming apparatus.

When the image forming apparatus is not provided with a feature amount analyzing unit, the image processing server 43 is provided with a feature amount analyzing unit having the same function as the above-described feature amount analyzing unit 26d. What is necessary is just to transfer data and obtain | require a characteristic quantity.

As described in detail above, according to the second embodiment, it is possible to perform image processing with higher accuracy than in the image processing apparatus in an external device such as the image processing server 43. The cost merit can be obtained by sharing and using the same external device among many image processing apparatuses.

Further, by performing totaling and statistical processing of each image processing in a large number of image processing apparatuses by an external apparatus, a more optimal combination of filter coefficients for the image processing apparatuses is obtained, and the individual image processing apparatuses can be used alone. It is also possible to improve the processing speed at the same time.

[0154]

As described in detail above, the image processing apparatus of the present invention extracts a filter coefficient template holding at least two or more sets of filter coefficients, and extracts two sets of filter coefficients from the filter coefficient templates and synthesizes them. And a filter synthesis processing unit that creates a synthesis filter coefficient by performing the processing. When the filter synthesis processing unit synthesizes each filter coefficient, the filter synthesis processing unit has a size that is larger than each size of the filter coefficients. In this configuration, a zero coefficient consisting of zero is used.

Therefore, in the above configuration, it is possible to create a synthetic filter coefficient for performing a variety of filtering processes from a limited number of filter coefficients. The image processing can be ensured by the processing, and the filter synthesis processing unit can simplify the filter synthesis and simplify the image processing by synthesizing using the zero coefficient.

As described above, the image processing system of the present invention is provided with the data input / output unit for communicating with the external device, and the control unit for communicating with the external device performs at least a part of the image processing function. An image processing system provided with an image processing device provided to perform processing by the external device, and an external device communicably connected to the image processing device, wherein the image processing device further includes an image processing device. An image input unit for inputting data and a feature amount analyzing unit for analyzing a feature amount of the image in the image data by a feature extraction method such as a histogram and a fast Fourier transform are provided at least. And a feature amount correlation processing unit that obtains a feature amount having a higher correlation than the plurality of feature amounts extracted by the feature amount analysis unit.

[0157] Therefore, in the above configuration, image processing with higher precision than image processing performed inside the image processing apparatus can be performed by an external device, so that there is an effect that image processing can be performed with higher precision. . Further, in the above configuration, there is an effect that the cost merit can be obtained by sharing the external device.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an image processing apparatus according to a first embodiment of the present invention.

FIG. 2 is a block diagram illustrating a digital copying machine to which the image processing apparatus is applied.

FIG. 3 is an explanatory diagram of a screen for designating a filter coefficient to be used in the image processing apparatus.

FIG. 4 is an explanatory diagram of a screen for designating switching between automatic selection of a filter coefficient to be combined and designation by a user in the image processing apparatus.

FIG. 5 is an explanatory diagram of a screen for switching output of a synthesis filter coefficient in the image processing apparatus.

FIG. 6 is an explanatory diagram of a screen for switching processing when the partial normalization coefficient becomes 0 in the image processing apparatus when combining filter coefficients.

FIG. 7 is an explanatory diagram showing a filter combining method of the image processing apparatus.

FIG. 8 is an explanatory diagram showing another filter synthesis method of the image processing apparatus.

FIG. 9 is an explanatory diagram showing still another filter synthesizing method of the image processing apparatus.

FIG. 10 is an explanatory diagram showing a relationship between a processing result when an FFT is used as a feature amount analysis process and a filter coefficient used for filter synthesis in the image processing apparatus;
(A) is a graph of a processing result, (b) is a graph showing high-pass filter characteristics, and (c) is a graph showing low-pass filter characteristics.

FIG. 11 is a graph illustrating a typical histogram for each image type in the image processing apparatus;
(A) shows a full-color image such as a photograph, (b) shows a black and white document centered on characters, and (c) shows a business document including a colored figure or graph.

FIG. 12 is a block diagram illustrating an image processing apparatus according to a second embodiment of the present invention.

FIG. 13 is an explanatory diagram showing a flow of data between the image processing apparatus and an image processing server as an external apparatus connected to the image processing apparatus according to the present invention;

FIG. 14 is an explanatory diagram illustrating an image processing method in the image processing apparatus.

FIG. 15 is an explanatory diagram showing a configuration of the image processing server.

FIG. 16 is an explanatory diagram illustrating an example of image processing in the image processing server.

FIG. 17 is an explanatory diagram showing another example of image processing in the image processing server.

FIG. 18 is an explanatory diagram showing a process of obtaining a character constant expression for incorporating a synthesis filter into a system by a conventional method.

[Explanation of symbols]

 26h filter coefficient template 26k filter synthesis processing unit

 ──────────────────────────────────────────────────続 き Continued on front page F term (reference) 5B057 AA11 BA02 BA26 CA01 CA08 CA12 CA16 CB01 CB08 CB12 CB16 CC02 CE03 CE05 CE06 CH09 CH11 CH18 DA16 DA17 DB02 DB06 DB09 DC22 DC23 5C077 LL17 PP03 PP46 PP49 PQ08 PQ19 PQ22 SS01

Claims (21)

[Claims] 1. A filter coefficient template that holds at least two or more sets of filter coefficients, and a filter synthesis processing unit that extracts and synthesizes two sets of filter coefficients from the filter coefficient template to create a synthesized filter coefficient. The filter synthesis processing unit is characterized in that, when synthesizing each filter coefficient, a zero coefficient having a size larger than each size of each of the filter coefficients and including all zeros is used. Image processing device. 2. The image processing apparatus according to claim 1, wherein the filter synthesis processing unit extracts three or more sets of filter coefficients from the filter coefficient template, and generates a second set of filter coefficients from the three or more sets of filter coefficients. An image processing apparatus comprising: a first synthesis processing unit that generates one synthesis filter coefficient, and generates a second synthesis filter coefficient using the first synthesis filter coefficient and the remaining filter coefficients. 3. The image processing apparatus according to claim 1, wherein a user input unit, such as a touch panel, for inputting an instruction from a user, and a filter synthesis processing unit which receives the instruction from the user input unit. An image processing apparatus comprising: a first selection unit that selects a set of filter coefficients to be combined by the first selection unit. 4. The image processing apparatus according to claim 1, further comprising: an image input unit for inputting image data; and a feature extraction method such as a histogram and a fast Fourier transform. The image processing apparatus includes a feature amount analysis unit that analyzes the feature amount of the image, and a synthesis filter buffer that holds the synthesis filter coefficient.
An image processing apparatus, comprising: a first specifying unit that specifies a set of filter coefficients to be combined. 5. The image processing apparatus according to claim 1, wherein a user input unit, such as a touch panel, for inputting an instruction from a user, and a filter synthesis processing unit that synthesizes the image according to the instruction from the user input unit. A first selection unit for selecting a set of filter coefficients; an image input unit for inputting image data; and a feature amount for analyzing an image feature amount in the image data by a feature extraction method such as a histogram or a fast Fourier transform. An analysis unit and a synthesis filter buffer for holding synthesis filter coefficients are provided.
A first specifying unit for specifying a set of filter coefficients to be combined, and switching between designating each set of filter coefficients to be combined by an input from a user or designating a result of a feature amount analysis of an image. And a selection input unit for selecting according to an instruction from a user input unit based on an input from a user. 6. The image processing apparatus according to claim 4, wherein an average value of density, a density histogram, and a fast Fourier based on a part or all of the image data as a feature of the image to be analyzed by the feature analyzer. An image processing apparatus comprising: a second specifying unit that specifies a set of filter coefficients to be combined using at least one feature amount of conversion. 7. The image processing apparatus according to claim 4, wherein a characteristic amount of the image analyzed by the characteristic amount analysis unit is calculated based on a filter coefficient used for synthesis according to the magnitude of the characteristic amount. An image processing apparatus comprising a changing unit for changing strength. 8. The image processing apparatus according to claim 7, wherein the filter synthesizing unit filters the filter coefficient corresponding to the feature amount equal to or smaller than a certain size with respect to the feature amount of the image analyzed by the feature amount analysis unit. An image processing apparatus comprising a second selection unit that is omitted from synthesis candidates. 9. The image processing apparatus according to claim 1, wherein: a template instruction history memory for storing a history of a combination specified as a set of filter coefficients to be synthesized; Based on the filter synthesis history stored in the template instruction history memory, a synthesis filter is created using the filter coefficients of frequently specified combinations, and is additionally stored in the filter coefficient template. An image processing apparatus comprising: a second synthesis processing unit that uses a corresponding additionally stored filter coefficient template when performing synthesis processing. 10. An image processing apparatus according to claim 1, wherein a filter coefficient output unit for displaying or outputting an output of a filter synthesis result as numerical information of a filter coefficient, and an image using the filter synthesis result. A filter processing unit for performing filter processing on the data, an output switching unit for switching between outputting and not outputting the filter synthesis result to the filter coefficient output unit and the filter processing unit; a filter coefficient output unit for filtering the filter synthesis result; An image processing apparatus, comprising: an output control unit that outputs to at least one of the processing units. 11. The image processing apparatus according to claim 1, wherein the sizes of the first filter coefficient and the second filter coefficient for synthesizing are (2i + 1) × (2j + 1) and (2k + 1) × ( 2l + 1) (where i,
j, k, l are positive integers) and the size is (2i + 2k + 1)
When generating a composite filter coefficient of × (2j + 2l + 1), a size (2i) in which the first filter coefficient is regarded as a pixel value
Using a + 4k + 1) × (2j + 4l + 1) image filter coefficient and a rotation filter coefficient obtained by rotating the second filter coefficient, a product-sum operation is performed between the image filter coefficient and the rotation filter coefficient. Center (2i + 2k +
1) An image processing apparatus including a third synthesis processing unit that uses a calculation result of × (2j + 2l + 1) as a synthesis filter coefficient. 12. The image processing apparatus according to claim 11, wherein the first filter coefficient is multiplied by a first constant in which all of the first filter coefficients are integers before generating the image filter coefficients, and Before creation, the second filter coefficient is multiplied by a second constant that makes all coefficients of the second filter coefficient an integer, and a product-sum operation of a rotation filter is performed on the image filter coefficient by an integer operation, and the center of the operation result (2i + 2k +
1) × (2j + 2l + 1) is calculated as a first constant and a second constant
An image processing apparatus comprising: a fourth synthesis processing unit that uses a result obtained by dividing by a product of constants as a synthesis filter coefficient. 13. The image processing apparatus according to claim 1, wherein the sizes of the first filter coefficient and the second filter coefficient for synthesizing are (2i + 1) × (2j + 1) and (2k + 1) × ( 2l + 1), the size (2i + 4k + 1) × where the first filter coefficient is regarded as a pixel value
Using an image filter coefficient of (2j + 4l + 1) and a rotation filter obtained by rotating the second filter coefficient, a partial filter is created based on the size of the image filter coefficient and the synthesis filter coefficient [(2m + 1) × (2n + 1)] when the rotation filter is created A normalization coefficient is obtained (where i, j, k, l, m, and n are positive integers), and each coefficient of the second filter coefficient is divided by a partial normalization coefficient, so that the size of the synthesis filter coefficient is 1
An image processing apparatus, comprising: a fifth synthesis processing unit that generates a (2m + 1) × (2n + 1) synthesis filter coefficient having a size independent of the sizes of the filter coefficient and the second filter coefficient. 14. The image processing apparatus according to claim 13, further comprising: a sixth synthesis processing unit that stops synthesis processing of a filter when 0 is obtained as a partial normalization coefficient. . 15. An image processing apparatus according to claim 13, wherein 0 is obtained as a partial normalization coefficient.
A seventh synthesis processing unit that generates a synthesis filter coefficient by the fourth synthesis processing unit according to claim 2 or the fifth synthesis processing unit according to claim 13, and creates a target filter coefficient of a size to be generated from the synthesis filter coefficient. An image processing apparatus comprising: 16. The image processing apparatus according to claim 11, wherein the method of synthesizing the filter coefficient is performed.
An image processing apparatus, comprising: at least two of any of the third to seventh synthesis processing units according to (1), and a third selection unit that selects a filter synthesis process. 17. A filter coefficient template holding at least two or more sets of filter coefficients is provided for image processing, and two sets of filter coefficients are extracted from the filter coefficient templates and combined to create a combined filter coefficient. The filter synthesis processing unit is provided for image processing such that when synthesizing each filter coefficient, a zero coefficient having a size larger than each size of each of the filter coefficients and all coefficients being zero is used. A data input / output unit for communicating with an external device having at least a part of image processing is provided, and a control unit that communicates with the external device via the data input / output unit has at least a part of image processing functions An image processing apparatus, which is provided so as to process the image by the external device. 18. An image processing system comprising: the image processing apparatus according to claim 17; and an external device communicably connected to the image processing apparatus. The image processing apparatus further comprises: An image input unit for performing input, and a feature amount analyzing unit for analyzing a feature amount of the image in the image data by a feature extraction method such as a histogram and a fast Fourier transform. An image processing system comprising: a feature amount correlation processing unit that obtains a feature amount having a higher correlation than a plurality of feature amounts extracted by a quantity analysis unit. 19. An image processing system according to claim 18, wherein said external device further comprises: a filter selecting unit for selecting a set of filter coefficients from a filter template based on said strongly correlated feature amount; An image processing system comprising: a new filter generation unit that generates a new filter coefficient when it is determined that there is no filter coefficient. 20. The image processing system according to claim 18, wherein an external device is provided with a filter synthesizing unit that synthesizes a filter based on the selected filter coefficient. . 21. An image processing system according to claim 18, 19, or 20, wherein the processing contents of the image processing apparatus are totaled, and an optimal combination of templates corresponding to the feature amounts extracted by said image processing apparatus is determined. An image processing system, wherein the statistical processing unit to be obtained is provided in an external device.