JP2004227406A - Image processor, program and storage medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processor capable of realizing the reduction of a system scale and a processing time in comparison of images. <P>SOLUTION: This image processor processing image data, which is a composite machine, extracts feature values from a first image and a second image each divided into a plurality of areas and compressed in each of the plurality of areas (S3), and compares the first image and the second image on the basis of the feature values extracted from the images in each of the plurality of areas (S6). Therefore, a process or the like for dividing the image into the plurality of areas is not required, and the image comparison can be executed with high accuracy by a simple process, so that the reduction of the system scale and the reduction of the processing time can be realized. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an image processing device, a program, and a storage medium.
[0002]
[Prior art]
2. Description of the Related Art In recent years, with the spread of image processing apparatuses such as scanners, digital cameras, computers, copiers, and multifunction peripherals (MFPs), digital image data is stored in a storage device such as a memory or a hard disk, or is stored on an optical disk such as a CD-ROM. It is familiar to memorize and further transmit via the Internet or the like. In particular, recently, computers, multifunction devices, and the like are often used as image servers. Various image processing is performed on such image data.
[0003]
For example, as a general image processing flow for a multifunction device or the like,
(1) Image input → tilt angle detection → tilt angle correction → area identification processing → character recognition processing
▲ 2 ▼ Image input → Tilt angle detection → Tilt angle correction → Title extraction → Use title as file name for document management
(3) Color image input → binarization processing → inclination angle detection (same as the processing of (1) or (2) thereafter)
(4) Color image input → Binarization processing (same as the processing of (3)) → Specify the character color of the character area or title area → Reproduce and display
And so on.
[0004]
The flow of image processing for searching for images (or image processing for classifying images) is as follows.
(5) Image input → feature extraction → search for images with similar features (subsequent classification)
Such processing is common. Here, the feature amount includes color, texture, edge information, and the like.
[0005]
Various techniques executed in such a flow of image processing have been proposed. For example, a technique of storing image data of a plurality of resolutions and selectively outputting an image of a resolution suitable for an output device (for example, see Patent Document 1), generating image data of a plurality of resolutions, and creating a ruled line from the image of each resolution (See, for example, Patent Document 2), a technique of generating image data of a plurality of resolutions, obtaining an optimum resolution based on a character type, and then performing a character recognition process (see, for example, Patent Document 3) Etc. have been proposed.
[0006]
On the other hand, recently, high-definition images can be easily obtained by various techniques, but the data size of such high-definition still images tends to be large, and handling of high-definition still images becomes difficult. ing. At present, JPEG (Joint Photographic Experts Group) and JPEG2000, an image compression / expansion method succeeding JPEG, which became an international standard in 2001, are used as image compression / expansion algorithms for facilitating the handling of such high-definition still images. The feature of the image data compressed by JPEG2000 is that it has images of a plurality of resolutions, one image is previously divided into a plurality of regions (tile regions) by a tiling process, and a frequency component (LL) Component, LH component, HL component, HH component).
[0007]
[Patent Document 1]
JP 2000-187725 A
[Patent Document 2]
JP 2001-127772 A
[Patent Document 3]
JP-A-2002-24766
[0008]
[Problems to be solved by the invention]
Conventional techniques generate image data of a plurality of resolutions, store the generated image data of a plurality of resolutions, divide an image into a plurality of regions, and obtain edge information which is a feature amount. Need to be Since it is necessary to execute various processes as described above, the system scale of the image processing increases, and the processing time of the image processing also increases.
[0009]
Further, in the conventional technique, since the feature amount is often extracted from a part of the image, the image comparison is not performed with high accuracy when searching or classifying the image, and as a result, the image search or classification is performed accurately. May not be done.
[0010]
An object of the present invention is to provide an image processing apparatus, a program, and a storage medium that can realize a reduction in system scale and processing time.
[0011]
An object of the present invention is to provide an image processing device, a program, and a storage medium that can execute image comparison with high accuracy.
[0012]
[Means for Solving the Problems]
An image processing apparatus according to claim 1, wherein a feature amount extracting unit extracts a feature amount for each of the plurality of regions from the first image and the second image which are divided into a plurality of regions and compressed. An image comparing the first image and the second image based on a feature amount extracted for each of a plurality of regions from the first image and the second image by the feature amount extracting unit; Comparing means.
[0013]
Therefore, a feature amount is extracted for each of the plurality of regions from the first image and the second image which are divided into a plurality of regions and compressed, and the first image and the second image are extracted based on the extracted feature amounts. By comparing the image with the image, processing for dividing the image into a plurality of regions is not required, and the image comparison can be performed with high accuracy by simple processing. As a result, the system scale is reduced and the processing is reduced. Time can be reduced.
[0014]
According to a second aspect of the present invention, in the image processing apparatus according to the first aspect, the feature amount extracting unit identifies a region attribute of the region and extracts the identified region attribute as a feature amount.
[0015]
Therefore, by identifying the region attribute of the region, the feature amount can be easily extracted, and the image comparison can be performed with high accuracy.
[0016]
According to a third aspect of the present invention, in the image processing apparatus of the second aspect, the first image and the second image are compressed by a procedure of two-dimensional wavelet transform, quantization, and encoding for each of a plurality of regions. The feature value extracting means identifies a region attribute of the region based on frequency component information included in a two-dimensional wavelet coefficient of the region.
[0017]
Therefore, by identifying the region attribute of the region based on the frequency component information included in the two-dimensional wavelet coefficient of the region, the process for extracting the feature amount is simplified, and the region attribute can be easily and accurately determined. The identification can be performed, and as a result, the processing time can be reduced.
[0018]
According to a fourth aspect of the present invention, in the image processing apparatus according to the third aspect, the feature amount extracting unit determines a region attribute of the tile region based on an HL component and an LH component included in a two-dimensional wavelet coefficient of the region. As a tablespace attribute.
[0019]
Therefore, by identifying the region attribute of the region as the table region attribute based on the HL component and the LH component included in the two-dimensional wavelet coefficient of the region, it is possible to easily and accurately identify the table region. .
[0020]
According to a fifth aspect of the present invention, in the image processing apparatus according to the first aspect, the feature amount extracting unit converts the ruled line information based on the HL component and the LH component included in the two-dimensional wavelet coefficient of the region. Extract as
[0021]
Therefore, by extracting the ruled line information based on the HL component and the LH component included in the two-dimensional wavelet coefficient of the area, it is possible to easily and accurately extract the ruled line information of the image.
[0022]
According to a sixth aspect of the present invention, in the image processing apparatus according to any one of the first to fifth aspects, an operation unit that receives an operation by an operator and a feature amount is extracted based on the operation of the operator on the operation unit. Area designating means for designating an area which needs to be performed, wherein the feature quantity extracting means extracts a feature quantity of only the area designated by the area designating means.
[0023]
Therefore, by specifying an area from which a feature amount needs to be extracted by an operation performed by the operator on the operation unit, the feature amount is extracted only from the specified area. The processing time can be shortened as compared with the case of performing the processing.
[0024]
According to a seventh aspect of the present invention, in the image processing apparatus according to any one of the first to sixth aspects, the image comparing unit is configured to determine each of the first image and the second image by the feature amount extracting unit. A feature amount comparing unit that compares a feature amount extracted for each of the plurality of regions for each of the plurality of regions, and matching between the feature amount of the first image and the feature amount of the second image by the feature amount comparing unit. A number-of-matches comparing means for comparing the number with a predetermined number of matches; and if the number of matches is determined by the number-of-matches comparing means to be greater than the predetermined number of matches, the first image and the second image are compared. Image specifying means for specifying an image having a high degree of coincidence.
[0025]
Therefore, by comparing the feature amount extracted for each of the plurality of regions from the first image and the second image for each of the plurality of regions, it is possible to easily and accurately perform the image comparison.
[0026]
According to an eighth aspect of the present invention, in the image processing apparatus according to the seventh aspect, the feature amount extracting unit extracts feature amounts of all the regions, and the feature amount comparing unit compares the feature amounts of all the regions. I do.
[0027]
Therefore, by comparing the feature values of all the regions, it is possible to perform the image comparison with higher accuracy than when comparing the feature values of a part of the regions.
[0028]
According to a ninth aspect of the present invention, in the image processing apparatus according to any one of the first to eighth aspects, a reading optical system for optically reading an image from a document is provided.
[0029]
Therefore, an image can be read from a document, and as a result, various processes such as image processing can be performed on the image.
[0030]
According to a tenth aspect of the present invention, in the image processing apparatus according to any one of the first to ninth aspects, the compressed image is decompressed for each of a plurality of regions, decompressed, and decompressed by two-dimensional wavelet inverse transform. Extending means for performing the extension.
[0031]
Therefore, the compressed image (code data) is expanded for each of a plurality of areas by decoding, inverse quantization, and inverse two-dimensional wavelet transform, whereby the compressed image is expanded. As a result, the expanded image is expanded. It is possible to execute the display of the displayed image on a display device or the like, the printing on paper or the like.
[0032]
According to an eleventh aspect of the present invention, in the image processing apparatus according to the tenth aspect, there is provided a printer engine for forming an image decompressed by the decompression means on a recording material.
[0033]
Therefore, it is possible to form an image of the expanded image on a recording material such as paper.
[0034]
The program according to claim 12, which is interpreted by a computer provided in the image processing apparatus, and instructs the computer to convert each of a plurality of regions from a first image and a second image which are divided into a plurality of regions and compressed. A feature value extraction function for extracting feature values to the first image and the second image based on feature values extracted for each of a plurality of regions by the feature value extraction function. And an image comparison function of comparing the image with the second image.
[0035]
Therefore, a feature amount is extracted for each of the plurality of regions from the first image and the second image which are divided into a plurality of regions and compressed, and the first image and the second image are extracted based on the extracted feature amounts. By comparing the image with the image, processing for dividing the image into a plurality of regions is not required, and the image comparison can be performed with high accuracy by simple processing. As a result, the system scale is reduced and the processing is reduced. Time can be reduced.
[0036]
According to a thirteenth aspect of the present invention, in the program according to the twelfth aspect, the feature amount extracting function identifies a region attribute of the region and extracts the identified region attribute as a feature amount.
[0037]
Therefore, by identifying the region attribute of the region, the feature amount can be easily extracted, and the image comparison can be performed with high accuracy.
[0038]
According to a fourteenth aspect of the present invention, in the program according to the thirteenth aspect, the first image and the second image are compressed by a procedure of two-dimensional wavelet transform, quantization, and encoding for each of a plurality of regions. The feature extraction function identifies a region attribute of a region based on frequency component information included in a two-dimensional wavelet coefficient of the region.
[0039]
Therefore, by identifying the region attribute of the region based on the frequency component information included in the two-dimensional wavelet coefficient of the region, the process for extracting the feature amount is simplified, and the region attribute can be easily and accurately determined. The identification can be performed, and as a result, the processing time can be reduced.
[0040]
According to a fifteenth aspect of the present invention, in the program according to the fourteenth aspect, the feature amount extracting function displays an area attribute of the tile area based on an HL component and an LH component included in a two-dimensional wavelet coefficient of the area. Identify as an area attribute.
[0041]
Therefore, by identifying the region attribute of the region as the table region attribute based on the HL component and the LH component included in the two-dimensional wavelet coefficient of the region, it is possible to easily and accurately identify the table region. .
[0042]
According to a sixteenth aspect of the present invention, in the program according to the twelfth aspect, the feature amount extracting function extracts ruled line information as a feature amount based on HL components and LH components included in a two-dimensional wavelet coefficient of the area. I do.
[0043]
Therefore, by extracting the ruled line information based on the HL component and the LH component included in the two-dimensional wavelet coefficient of the area, it is possible to easily and accurately extract the ruled line information of the image.
[0044]
According to a seventeenth aspect of the present invention, in the program according to any one of the twelfth to thirteenth aspects, the program has an area designating function for designating an area in which a feature amount needs to be extracted based on an operation performed by an operator on an operation unit. The program is executed by a computer, and the feature amount extracting function extracts a feature amount of only an area specified by the area specifying function.
[0045]
Therefore, by specifying an area for which a feature amount needs to be extracted by an operation on the operation unit, the feature amount is extracted only from the specified area, and the feature amount is extracted from all the areas. The processing time can be shortened as compared with the case of performing the processing.
[0046]
The invention according to claim 18 is the program according to any one of claims 12 to 17, wherein the image comparison function uses the feature amount extraction function to output a plurality of images from the first image and the second image. A feature value comparison function of comparing the feature value extracted for each region for each of a plurality of regions; and a number of matches between the feature value of the first image and the feature value of the second image by the feature value comparison function. A coincidence number comparison function for comparing with a predetermined number of coincidences, and when the coincidence number comparison function determines that the number of coincidences is larger than the predetermined number of coincidences, the degree of coincidence between the first image and the second image is determined. And causing the computer to execute an image specifying function of specifying an image having a high score.
[0047]
Therefore, by comparing the feature amount extracted for each of the plurality of regions from the first image and the second image for each of the plurality of regions, it is possible to easily and accurately perform the image comparison.
[0048]
According to a nineteenth aspect of the present invention, in the program according to the eighteenth aspect, the feature amount extracting function extracts feature amounts of all the regions, and the feature amount comparing function compares the feature amounts of all the regions.
[0049]
Therefore, by comparing the feature values of all the regions, it is possible to perform the image comparison with higher accuracy than when comparing the feature values of a part of the regions.
[0050]
According to a twentieth aspect of the present invention, there is provided the program according to any one of the twelfth to nineteenth aspects, wherein the compressed image is expanded for each of a plurality of regions by a procedure of decoding, inverse quantization, and inverse two-dimensional wavelet transform. Causing the computer to execute a function.
[0051]
Therefore, the compressed image (code data) is expanded by decoding, dequantization, and inverse two-dimensional wavelet transform for each of a plurality of regions, so that the compressed image is expanded. As a result, the expanded image is expanded. It is possible to execute the display of the displayed image on a display device or the like, the printing on paper or the like.
[0052]
A computer-readable storage medium according to a twenty-first aspect stores the program according to any one of the twelfth to twentieth aspects.
[0053]
Therefore, the same operation as the invention according to any one of claims 12 to 20 is achieved.
[0054]
BEST MODE FOR CARRYING OUT THE INVENTION
A first embodiment of the present invention will be described with reference to FIGS.
[0055]
In the present embodiment, the “JPEG2000 algorithm” is used. However, since the JPEG2000 algorithm itself is well known in various documents and gazettes, the details are omitted, and the outline is described.
[0056]
FIG. 1 is a functional block diagram for explaining the outline of the JPEG2000 algorithm. The JPEG2000 algorithm includes a color space conversion / inverse conversion unit 100, a two-dimensional wavelet conversion / inverse conversion unit 101, a quantization / inverse quantization unit 102, an entropy encoding / decoding unit 103, and a tag processing unit 104. I have.
[0057]
One of the features of JPEG2000 is that a two-dimensional discrete wavelet transform (DWT: Discrete Wavelet Transform) having an advantage of high image quality in a high compression area is used. Another major feature is that a functional block called a tag processing unit 104 for performing code formation is added at the last stage, and a code stream as code string data is generated and interpreted. JPEG2000 can realize various convenient functions by the code stream.
[0058]
Note that a color space conversion / inversion unit 100 is often provided in the input / output part of the image. The color space conversion / inverse conversion unit 100 includes, for example, an RGB color system including R (red) / G (green) / B (blue) components of a primary color system, and Y (yellow) / M of a complementary color system. This is a part that performs conversion from the YMC color system composed of each component of (magenta) / C (cyan) to YCrCb or the YUV color system or vice versa.
[0059]
Hereinafter, the JPEG2000 algorithm, particularly the wavelet transform, will be described.
[0060]
FIG. 2 is a schematic diagram schematically illustrating an example of each of the divided components of an original image that is a color image. In a color image, as shown in FIG. 2, each component 110 of the image is generally separated by, for example, an RGB primary color system. Further, each component 110 of the image is divided by tiles 111 which are rectangular areas. Each tile 111, for example, R00, R01, ..., R15 / G00, G01, ..., G15 / B00, B01, ..., B15 constitutes a basic unit when executing the compression / decompression process. Therefore, the compression / expansion operation is performed independently for each component 110 and for each tile 111.
[0061]
At the time of encoding image data (see FIG. 1), data of each tile 111 of each component 110 is input to the color space conversion / inverse conversion unit 100, subjected to color space conversion, and then subjected to two-dimensional wavelet conversion / inverse conversion. The unit 101 applies a two-dimensional wavelet transform (forward transform) and spatially divides the frequency band.
[0062]
FIG. 3 is a schematic diagram schematically showing subbands at each decomposition level when the number of decomposition levels is three. The two-dimensional wavelet transform / inverse transform unit 101 performs a two-dimensional wavelet transform on the tile image (decomposition level 0: 0LL) obtained by dividing the image into tiles, and performs a subband (1LL) indicated at the decomposition level 1 , 1HL, 1LH, 1HH). Subsequently, the two-dimensional wavelet transform / inverse transform unit 101 performs a two-dimensional wavelet transform on the low-frequency component 1LL in this layer, and separates the subbands (2LL, 2HL, 2LH, 2HH) shown in the decomposition level 2. I do. Then, the two-dimensional wavelet transform / inverse transform unit 101 similarly performs the two-dimensional wavelet transform on the low-frequency component 2LL in the same manner, and the sub-bands (3LL, 3HL, 3LH, 3HH) shown in the decomposition level 3 To separate. In FIG. 3, the subbands to be encoded at each decomposition level are shown in gray. For example, if the number of decomposition levels is 3, the gray subbands (3HL, 3LH, 3HH, 2HL, 2LH, 2HH, 1HL, 1LH, 1HH) are to be encoded, and the 3LL subband is not encoded. .
[0063]
Next, in the quantization / dequantization unit 102 (see FIG. 1), after bits to be encoded are determined in the specified encoding order, a context is generated from bits around the target bits. The wavelet coefficients after the quantization process are divided into non-overlapping rectangles called “precincts” for each subband. This was introduced to make efficient use of memory in the implementation. Furthermore, each precinct is divided into non-overlapping rectangular “code blocks”. This is a basic unit when performing entropy coding.
[0064]
Although the coefficient values after the wavelet transform can be quantized and encoded as they are, in JPEG2000, in order to increase the encoding efficiency, the coefficient values are decomposed into "bit planes", and each pixel or code block is decomposed. The bit planes can be prioritized.
[0065]
The entropy coding / decoding unit 103 (see FIG. 1) performs coding on each tile 111 of each component 110 by probability estimation from the context and the target bit. Thus, the encoding process is performed on all the components 110 of the image in units of the tiles 111.
[0066]
Finally, the tag processing unit 104 (see FIG. 1) combines all coded data from the entropy coding / decoding unit 103 into one code stream (code string data) and adds a tag to the code stream. I do. Here, FIG. 4 is a schematic diagram schematically showing an example of the structure of the code stream. Tag information called a header (main header (Main header), tile part header) is added to the beginning of the code stream and the beginning of the partial tiles constituting each tile 111. Coded data (bit stream). Then, at the end of the code stream, tag information (end of codestream) is added again.
[0067]
On the other hand, at the time of decoding, image data is generated from the code stream of each tile 111 of each component 110, contrary to the time of encoding. In this case, as shown in FIG. 1, the tag processing unit 104 interprets the tag information added to the code stream (code string data) input from the outside, and converts the code stream into the code of each tile 111 of each component 110. The stream is decomposed into streams, and decoding processing is performed for each code stream of each tile 111 of each component 110. At this time, the position of the bit to be decoded is determined in the order based on the tag information in the code stream, and the quantization / inverse quantization unit 102 sets the peripheral bits of the target bit position (decoding is already completed. Is generated from the sequence of Then, the entropy coding / decoding unit 103 generates a target bit by performing decoding by probability estimation from the context and the code stream, and writes the target bit at the position of the target bit. Since the data decoded in this manner is spatially divided for each frequency band, the two-dimensional wavelet transform / inverse transform unit 101 performs an inverse two-dimensional wavelet transform on each of the components, thereby obtaining each component in the image data. Each tile 111 at 110 is restored. The restored data is converted by the color space conversion / inversion unit 100 into the original color system data.
[0068]
Next, a configuration example of the multifunction peripheral 1 which is the image processing apparatus of the present embodiment will be described. The multifunction device 1 according to the present embodiment has a multifunction such as a copy function, a printer function, a scanner function, a facsimile function, and an image server function.
[0069]
FIG. 5 is a longitudinal sectional view schematically showing the multifunction peripheral 1 of the present embodiment. The MFP 1 includes a scanner 2 that is an image reading unit that reads a document image from a document, and a printer 3 that is an image forming unit that forms an image read by the scanner 2 on a recording material such as paper.
[0070]
A contact glass 5 on which a document (not shown) is placed is provided on the upper surface of the main body case 4 of the scanner 2. The original is placed with the original surface facing the contact glass 5. On the upper side of the contact glass 5, a document pressure plate 6 (which may be a so-called ADF) for pressing a document placed on the contact glass 5 is provided.
[0071]
Below the contact glass 5, a reading optical system 7 for optically reading a document image is provided. The reading optical system 7 includes a first traveling body 10 on which a light source 8 for emitting light and a mirror 9 are mounted, a second traveling body 13 on which two mirrors 11 and 12 are mounted, and a mirror 9 via an imaging lens 14. , 11, and 12 are configured by a CCD (Charge Coupled Device) image sensor 15 and the like that receive the light guided by the light. The CCD image sensor 15 functions as a photoelectric conversion element that generates photoelectric conversion data by photoelectrically converting reflected light from a document imaged on the CCD image sensor 15. The photoelectric conversion data is a voltage value having a magnitude corresponding to the intensity of light reflected from the document. The first and second traveling bodies 10 and 13 are provided so as to be able to reciprocate along the contact glass 5, and at the time of reading an original image, which will be described later, have a 2: 1 speed by a moving device such as a motor (not shown). Scanning travel in the sub-scanning direction by the ratio. As a result, exposure scanning of the document reading area by the reading optical system 7 is performed. In the present embodiment, the reading optical system 7 side is shown as a fixed document type that performs scanning, but the reading optical system 7 side may be a document moving type in which the position is fixed and the document side moves.
[0072]
The printer 3 includes a recording material path 20 from a recording material holding unit 16 for holding a recording material such as a sheet of paper to an output unit 19 via an electrophotographic printer engine 17 and a fixing unit 18.
[0073]
The printer engine 17 uses a photoconductor 21, a charger 22, an exposure unit 23, a developing unit 24, a transfer unit 25, a cleaner 26, and the like to form a toner image formed around the photoconductor 21 by electrophotography on a recording material. The transferred and transferred toner image is fixed on the recording material by the fixing device 18. In the present embodiment, the printer engine 17 forms an image by an electrophotographic method. However, the present invention is not limited to this. For example, various image forming methods such as an ink jet method, a sublimation type thermal transfer method, and a direct thermal recording method may be used. May be used to form an image.
[0074]
The MFP 1 is controlled by a control system including a plurality of microcomputers. FIG. 6 is a block diagram schematically showing an electrical connection of a control system related to image processing among these control systems. The control system includes a CPU 30, a ROM 31, a RAM 32, an operation panel 33, an IPU (Image Processing Unit) 34, an I / O port 35, a communication control unit 36, and the like, which are connected by a bus 37. The CPU 30 performs various calculations and centrally controls processing such as image processing. The ROM 31 stores various programs and fixed data related to processing executed by the CPU 30. The RAM 32 functions as a work area for the CPU 30 and also functions as a memory for temporarily storing image data (for example, image files). The operation panel 33 is provided with a display such as an LCD (Liquid Crystal Display), a plurality of operation keys including a hard key, a touch panel, and the like (all not shown). Functions as an operation unit. The IPU 34 includes hardware related to various types of image processing, and the ROM 31 includes a nonvolatile memory such as an EEPROM or a flash memory, and also functions as a memory for storing image data (for example, an image file). Here, the program stored in the ROM 31 can be rewritten as a program downloaded from an external device (not shown) via the I / O port 35 under the control of the CPU 30. In the present embodiment, the ROM 31 functions as a storage medium for storing a program. The communication control unit 36 has a function of transmitting and receiving data between the multifunction device 1 and an external device (not shown) via a network or the like, and is used to connect to a facsimile modem function and a public telephone line network. Network control function, LAN (Local Area Network) control function, and the like.
[0075]
Next, an outline of image processing in the multifunction peripheral 1 of the present embodiment will be described with reference to FIG. FIG. 7 is a functional block diagram for explaining an outline of image processing in the multifunction peripheral 1.
[0076]
The image processing of the multifunction device 1 is performed by using a compression unit 40 composed of each functional block described with reference to FIG. 1 and a first image and a second image which are divided into a plurality of regions and compressed. The image processing apparatus includes a feature amount extraction unit 41 that extracts a feature amount for each of the plurality of regions, and an image comparison unit 42 that compares the first image and the second image based on the extracted feature amounts for each of the plurality of regions. ing.
[0077]
Basically, image data of an image read by the scanner 2 is compression-encoded based on the JPEG2000 algorithm to generate a code stream of the image. That is, the image is divided into a plurality of tile areas (rectangular areas) as one or a plurality of areas, and pixel values are discrete wavelet-transformed for each of the tile areas and compression-coded hierarchically. Thereafter, a feature amount is extracted for each tile region from the compressed image, and the original image (first image) and another image (second image) are compared based on the extracted feature amounts. Thereby, it is possible to search for a similar image of the original image or the same image that matches the original image, and to classify those images. Here, the plurality of regions is not limited to the tile region. Note that the image data compressed based on the JPEG2000 algorithm may be received, and the received image may be subjected to processing by the feature amount extraction unit 41 and the image comparison unit 42.
[0078]
The functions of the compression unit 40, the feature amount extraction unit 41, the image comparison unit 42, and the like are executed by image processing performed by the CPU 30 based on a program stored in the ROM 31, but are not limited thereto. Instead, it may be executed by, for example, image processing performed by hardware by the IPU 34.
[0079]
Here, a method of extracting a feature amount in the feature amount extraction unit 41 will be described. The feature amount extraction unit 41 executes a region identification process for identifying a region attribute for each of a plurality of tile regions, and extracts the region attribute as a feature amount. The area attributes include, for example, various area attributes such as a character area, a photograph area, a figure area, a table area, a solid black area, a background area, and an indistinguishable area. Note that an unidentifiable area is an area whose area attribute could not be identified by the area identification processing. Further, the background area is a blank area corresponding to the margin of the document, but is not limited to this. For example, the background area may be an area obtained by adding a space between lines to the blank area.
[0080]
Next, image processing executed by the CPU 30 of the MFP 1 based on a program will be described. Here, image processing in the case of storing image data in order to use the MFP 1 as an image server will be described with reference to FIGS. FIG. 8 is a flowchart schematically showing the flow of image processing according to the present embodiment, FIG. 9 is an explanatory diagram schematically showing an example of an image divided into a plurality of tile areas, and FIG. 10 is a flow of feature amount extraction processing. 11 is a flowchart schematically showing the flow of the image comparison process, FIG. 12 is an explanatory diagram schematically showing the image comparison process, and FIG. 13 is a description schematically showing the result of the image comparison process. FIG.
[0081]
As shown in FIG. 8, first, the process waits for reading of a document image by the scanner 2 (N in step S1). When the operator opens the document pressing plate 6 of the scanner 2 and sets a document on the contact glass 5, closes the document pressing plate 6 and presses the start key of the operation panel 33, the scanner 2 performs a scanning operation of the reading optical system 7. An original image (first image) is read from a document set on the contact glass 5.
[0082]
When an original image is read from a document by the scanner 2 (Y in S1), the read original image is compressed based on the JPEG2000 algorithm (S2). That is, the original image is divided into a plurality of tile areas, and is hierarchically compressed in a procedure of two-dimensional wavelet transform, quantization, and encoding for each of the plurality of tile areas. As a result, the original image is divided into a plurality of tile areas. For example, as shown in FIG. 9, the original image is divided into a plurality of 4 × 6 = 24 tile areas.
[0083]
A feature amount extraction process for extracting a feature amount for each tile region from the compressed original image is executed (S3). Here, a feature amount extracting means or a feature amount extracting function is executed. Here, region identification processing for identifying a region attribute for each tile region is performed on the compressed original image, and the region attribute is extracted as a feature amount for each tile region.
[0084]
Here, the feature amount extraction processing will be described. As shown in FIG. 10, one tile area is selected from all the tile areas in the compressed original image (S11). A region identification process is performed on the selected tile region, and a region attribute of the tile region is extracted as a feature amount (S12). As a method of the area identification processing, a conventional method, for example, a method of identifying an area using the density of black runs is sufficient, and the method is well-known. Steps S11 to S12 are repeated until the area attributes of all tile areas are extracted (N in S13). As a result, the area attributes of all tile areas are identified as a character area, a photograph area, a figure area, a table area, a background area, an unidentifiable area, and the like, and are extracted for each tile area.
[0085]
Next, as shown in FIG. 8, it is determined whether or not the image (image data) is stored in the ROM 31 (S4). If it is determined that the image is not stored in the ROM 31 (N in S4), the compressed original image (image data) is stored in the ROM 31 (EEPROM, flash memory, or the like) (S5). Note that the region attribute, which is the feature amount extracted in step S3, is stored in the image data of the original image. However, the present invention is not limited to this. For example, the region attribute is stored as the region attribute data associated with the original image. May be.
[0086]
If it is determined that the image is stored in the ROM 31 (Y in S4), the original image compressed in step S2 is stored in the ROM 31 based on the area attribute extracted for each tile area in step S3. An image comparison process is performed for comparison with the image (second image) (S6). Here, an image comparison unit or an image comparison function is executed.
[0087]
Here, the image comparison processing will be described. As shown in FIG. 11, one tile area is selected from all the tile areas (S21: see FIG. 12), and the area attributes of each selected tile area are compared (S22: see FIG. 12). Here, a feature amount comparison unit or a feature amount comparison function is executed. For example, in FIG. 12, the area attributes of the tile area indicated by number 1 are compared between the image A and the image B. Thus, each selected tile area is a tile area located at the same position in the image. Next, steps S21 to S22 are repeated until the area attributes of all tile areas are compared (N in S23). Thus, the region attributes are compared for each tile region, and it is determined whether or not the region attributes of the tile regions match (see FIG. 13). In FIG. 13, １３ indicates that the region attributes match, and X indicates that the region attributes do not match.
[0088]
In this way, the area attributes of all tile areas are compared, and then the number of matches, ie, the number of 〇, is compared with a predetermined number of matches. Here, the matching number comparing means or the matching number comparing function is executed. If it is determined that the number of 〇 is greater than the predetermined number of matches, the compressed original image and the image stored in the ROM 31 are specified as images with a high degree of matching (similar images, identical images, etc.). Here, the image specifying means or the image specifying function is executed. For example, in FIG. 12, the predetermined number of matches is set to 20, and if the number of 〇 is greater than 20 or more than 20, the compressed original image and the image stored in the ROM 31 are regarded as similar images. When the number of matches is set to 24 and the number of 〇 is 24, the compressed original image and the image stored in the ROM 31 are specified as the same image. The predetermined number of matches is set before the multifunction device 1 is shipped from the factory. However, the present invention is not limited to this. For example, the predetermined number of matches may be set by an operation on the operation panel 33 by an operator.
[0089]
Thereafter, as shown in FIG. 8, the compressed original image (image data) is stored in the ROM 31 (EEPROM, flash memory, or the like) in association with the image having a high degree of coincidence (S5). When there is no image having a high degree of coincidence, the original image is stored in the ROM 31 without performing the association. Here, when a plurality of original images are read, the plurality of images are classified based on a region attribute which is a feature amount, and are stored in the ROM 31 in association with each similar image or the same image. As a result, an image file (dictionary file) composed of a plurality of classified images is formed. Note that an image compressed based on the JPEG2000 algorithm may be received by the communication control unit 36, and the received image may be subjected to image processing from step S3 to step S6.
[0090]
The image stored in the ROM 31 in this manner is expanded based on, for example, the JPEG2000 algorithm, and is formed on a recording material such as a sheet by the printer engine 17 or transmitted to an external device via a network or the like by the communication control unit 36. Or be done. Here, the decompression means or decompression function is executed.
[0091]
Also, when searching for a similar image or the same image with respect to an image selected by an operation on the operation panel 33 by the operator, the above-described image processing is similarly executed. As a result, images having a high degree of coincidence are retrieved, and similar images, identical images, and the like are obtained.
[0092]
As described above, in the present embodiment, the area of the tile area is determined for each tile area from the original image (first image) compressed based on the JPEG2000 algorithm and the image (second image) stored in the ROM 31. The attribute is identified, the area attribute of the identified tile area is extracted as a feature amount (S3), and the original image is compared with the image stored in the ROM 31 based on the extracted area attribute of each tile area. This makes it possible to execute various image processing utilizing the features of JPEG2000, and does not require a process of dividing an image into a plurality of regions or a process of forming an image with a plurality of resolutions. It is possible to perform image comparison with high accuracy in the processing, and as a result, it is possible to reduce the system scale and the processing time.
[0093]
Further, by extracting the region attributes of all the tile regions and comparing the region attributes of all the tile regions, it is possible to perform the image comparison with higher accuracy than when comparing the region attributes of a part of the tile regions. .
[0094]
Here, as the area attribute identification method in step S12, a wavelet feature can be used in addition to the conventional identification method. For example, a two-dimensional wavelet transform is performed on an image for each tile area (see FIG. 3). There is a method of identifying a region attribute of a tile region based on frequency component information (LL component, LH component, HL component, HH component) included in a two-dimensional wavelet coefficient of the tile region. Note that the LL component is a horizontal low frequency / vertical low frequency component, the LH component is a horizontal low frequency / vertical high frequency component, and the HL component is a horizontal high frequency / vertical low frequency component. , HH components are a horizontal high frequency component and a vertical high frequency component. In the JPEG2000 two-dimensional wavelet transform, a two-dimensional wavelet coefficient is calculated by applying a one-dimensional filter in a horizontal direction and a vertical direction, and the calculated two-dimensional wavelet coefficient is combined with a low-pass filter and a high-pass filter, as described above. Into four coefficient groups (LL component, LH component, HL component, and HH component).
[0095]
In such a method for identifying the region attribute based on the frequency component information, the region attribute of the tile region where the region of the HH component is widely distributed is determined by utilizing the fact that the HH component is high in the photograph region of the image. Identify as an attribute. Also, by utilizing the fact that the HH component is adjacent to and continuous in the character line direction in the character region of the image, the region attribute of the tile region in which the HH component is adjacent and continuous in the character line direction is identified as the character region attribute. I do. Furthermore, by utilizing the fact that the ruled lines of the table react to the HL component and the LH component, the intersection information of the ruled lines, which is the ruled line information, is extracted based on the HL and LH components. Identify the attribute as a tablespace attribute. In addition, taking advantage of the fact that HH, HL, and LH components other than the LL component are not so high in the background region, the region attribute of the tile region in which the HH, HL, and LH components are not so high as compared with other regions is changed to the background region attribute. Identify as Further, it is also possible to extract the inclination information (the amount of inclination) of the ruled line, which is the ruled line information, based on the HL and LH components, obtain the inclination of the image from the extracted inclination information of the ruled line, and detect the inclination angle.
[0096]
By using the frequency component information in this way, it is possible to execute the area identification with higher accuracy than the conventional area identification method. Further, conventionally, the ruled line information is extracted after obtaining the edge information. However, it is necessary to obtain the edge information by using the frequency component information of the HL component and the LH component of the edge information included in the compressed image. Therefore, the system scale and the processing time can be reduced.
[0097]
A second embodiment of the present invention will be described. This embodiment has the same basic configuration as that of the first embodiment, except for a feature amount extraction process executed by the CPU 30 of the MFP 1 based on a program. The point is that a tile area from which a feature amount needs to be extracted is specified based on a necessary area selected by an operator's operation on. The same parts as those described in the first embodiment are denoted by the same reference numerals, and description thereof will be omitted.
[0098]
A feature amount extraction process, which is a part of the image processing executed by the CPU 30 of the MFP 1 based on the program according to the present embodiment, will be described with reference to FIGS. FIG. 14 is a flowchart schematically showing the flow of the feature amount extraction processing according to the present embodiment, and FIG. 15 is an explanatory diagram schematically showing an example of a necessary area.
[0099]
As shown in FIG. 14, based on the required area selected by the operation on the operation panel 33 by the operator, a tile area for which an area attribute as a feature amount needs to be extracted is specified (S31). Here, the area specifying means or the area specifying function is executed. Here, numbers from 1 to 24 are assigned to the original image for each tile area (see FIG. 15). For example, when the operator operates the operation panel 33 to select a number from 9 to 20, a plurality of tile areas from 9 to 20 are selected as necessary areas (see FIG. 15). If the required area is not selected, for example, all tile areas 1 to 24 are selected as the required area.
[0100]
One tile area is selected from the designated plurality of tile areas (S32). A region identification process is performed on the selected tile region, and a region attribute of the tile region is extracted as a feature amount (S33). Steps S32 to S33 are repeated until the area attributes of the specified plurality of tile areas are extracted (N in S34). As a result, the area attributes of all tile areas are identified as a character area, a photograph area, a figure area, a table area, a background area, an unidentifiable area, and the like, and are extracted for each tile area.
[0101]
As described above, in the present embodiment, an effect similar to that of the first embodiment is obtained, and a tile area for which an area attribute needs to be extracted is specified by an operation on the operation panel 33 by the operator. Therefore, since the area identification processing is executed only for the designated tile area, the processing time can be reduced as compared with the case where the area identification processing is executed for all tile areas.
[0102]
In each of the embodiments, the MFP 1 is used as an image processing apparatus. However, the present invention is not limited to this. For example, a personal computer or the like may be used. In this case, the personal computer includes a CPU, a ROM, a RAM, an HDD (Hard Disk Drive) storing various programs, a CD-ROM drive, a communication control device for transmitting information by communication with an external device via a network, and a processing unit. A display device for displaying the progress and results to the operator, and an input device such as a keyboard and a mouse are provided. Here, the HDD functions as a storage medium for storing a program including the above-described JPEG2000 format compression / decompression program.
[0103]
Generally, a program installed in the HDD of a personal computer is recorded on an optical information recording medium such as a CD-ROM or a DVD-ROM, or a magnetic medium such as an FD, and the recorded program is stored in the HDD. Installed. Therefore, a portable storage medium such as an optical information recording medium such as a CD-ROM or a magnetic medium such as an FD can be a storage medium for storing a program including the above-described JPEG2000 format compression / decompression program. . Further, such a program may be fetched from outside via a communication control device, for example, and installed in the HDD.
[0104]
【The invention's effect】
According to the image processing apparatus of the present invention, a feature amount extracting means for extracting a feature amount for each of a plurality of regions from the first image and the second image which are divided into a plurality of regions and compressed. And comparing the first image and the second image based on the feature amount extracted for each of the plurality of regions from the first image and the second image by the feature amount extracting unit. And an image comparison unit that performs the image comparison, so that it is not necessary to perform a process of dividing an image into a plurality of regions, and it is possible to accurately perform an image comparison with a simple process. Reduction in size and processing time can be realized.
[0105]
According to the second aspect of the present invention, in the image processing apparatus according to the first aspect, the feature amount extracting means identifies a region attribute of a region and extracts the identified region attribute as a feature amount. The feature amount can be extracted, and the image comparison can be performed with high accuracy.
[0106]
According to the third aspect of the present invention, in the image processing apparatus according to the second aspect, the first image and the second image are subjected to two-dimensional wavelet transform, quantization, and encoding for each of a plurality of regions. Since the feature amount extracting means identifies the region attribute of the region based on the frequency component information included in the two-dimensional wavelet coefficients of the region, a process for extracting the feature amount is performed. Area attributes can be easily and simply identified with high accuracy, and as a result, processing time can be reduced.
[0107]
According to a fourth aspect of the present invention, in the image processing apparatus according to the third aspect, the feature amount extracting unit determines a tile region based on an HL component and an LH component included in a two-dimensional wavelet coefficient of the region. Since the area attribute is identified as the table area attribute, the table area can be easily identified with high accuracy.
[0108]
According to the fifth aspect of the present invention, in the image processing apparatus according to the first aspect, the feature amount extracting unit generates ruled line information based on HL components and LH components included in a two-dimensional wavelet coefficient of a region. Since the information is extracted as the feature amount, the ruled line information of the image can be easily extracted with high accuracy.
[0109]
According to a sixth aspect of the present invention, in the image processing apparatus according to any one of the first to fifth aspects, the operation unit for receiving an operation by the operator and the feature amount based on the operation of the operator on the operation unit. Area designating means for designating an area that needs to be extracted.The feature quantity extracting means extracts the feature quantity of only the area designated by the area designating means. By specifying an area from which a feature amount needs to be extracted by an operation, the feature amount is extracted only from the specified area. Therefore, the processing time is reduced as compared with the case where the feature amount is extracted from all the areas. Can be shortened.
[0110]
According to a seventh aspect of the present invention, in the image processing apparatus according to any one of the first to sixth aspects, the image comparing unit is configured to determine whether the first image and the second image are transmitted by the feature amount extracting unit. A feature amount comparing unit that compares a feature amount extracted for each of a plurality of regions from each of the plurality of regions, and a feature amount of the first image and a feature amount of the second image by the feature amount comparing unit. A match number comparing means for comparing the number of matches with a predetermined number of matches, and when the number of matches is determined to be larger than the predetermined number of matches, the first image and the second image And an image specifying unit for specifying as a high-matching image, a feature amount extracted for each of a plurality of regions from the first image and the second image is compared for each of a plurality of regions. Performing accurate and easy image comparisons It can be.
[0111]
According to an eighth aspect of the present invention, in the image processing apparatus according to the seventh aspect, the feature amount extracting unit extracts feature amounts of all regions, and the feature amount comparing unit includes feature amounts of all regions. Are compared with each other, image comparison can be performed with higher accuracy than when comparing the feature amounts of a part of the regions.
[0112]
According to the ninth aspect of the present invention, the image processing apparatus according to any one of the first to eighth aspects includes a reading optical system that optically reads an image from a document, so that the image can be read from the document. As a result, various processes such as image processing can be performed on this image.
[0113]
According to a tenth aspect of the present invention, in the image processing apparatus according to any one of the first to ninth aspects, a procedure of decoding a compressed image for each of a plurality of regions, inverse quantization, and inverse two-dimensional wavelet transform. Thus, the compressed image is decompressed, and as a result, the decompressed image can be displayed on a display device or the like or printed on paper or the like.
[0114]
According to the eleventh aspect of the present invention, in the image processing apparatus of the tenth aspect, a printer engine for forming an image decompressed by the decompression means on a recording material is provided. An image can be formed on a recording material.
[0115]
According to the program of the twelfth aspect of the present invention, the image processing apparatus is interpreted by a computer provided with the image processing apparatus, and instructs the computer to convert each of the first image and the second image divided into a plurality of regions and compressed into a plurality of regions. A feature amount extraction function of extracting a feature amount for each region; and a feature amount extraction function based on feature amounts extracted for each of a plurality of regions from the first image and the second image by the feature amount extraction function. Since the image comparison function of comparing the first image and the second image is executed, a process of dividing the image into a plurality of regions is not required, and the image comparison is executed accurately with a simple process. As a result, it is possible to reduce the system scale and the processing time.
[0116]
According to a thirteenth aspect of the present invention, in the program according to the twelfth aspect, the feature amount extracting function identifies a region attribute of a region and extracts the identified region attribute as a feature amount. Can be extracted, and image comparison can be performed with high accuracy.
[0117]
According to a fourteenth aspect of the present invention, in the program according to the thirteenth aspect, the first image and the second image are compressed by a procedure of two-dimensional wavelet transform, quantization, and encoding for each of a plurality of regions. The feature amount extraction function identifies the region attribute of the region based on the frequency component information included in the two-dimensional wavelet coefficient of the region, so that the process for extracting the feature amount is performed. The area attributes can be easily and simply identified with high accuracy, and as a result, the processing time can be reduced.
[0118]
According to a fifteenth aspect of the present invention, in the program according to the fourteenth aspect, the feature amount extracting function is configured to determine an attribute of the tile area based on an HL component and an LH component included in a two-dimensional wavelet coefficient of the area. Is identified as a table region attribute, the table region can be easily and accurately identified.
[0119]
According to a sixteenth aspect of the present invention, in the program according to the twelfth aspect, the feature amount extracting function converts the ruled line information based on the HL component and the LH component included in the two-dimensional wavelet coefficient of the region. , The ruled line information of the image can be easily extracted with high accuracy.
[0120]
According to a seventeenth aspect of the present invention, in the program according to any one of the twelfth to twelfth aspects, an area designating function for designating an area in which a feature amount needs to be extracted based on an operation performed by an operator on an operation unit. Is executed by the computer, and the feature amount extraction function extracts the feature amount only for the area designated by the area designation function. Therefore, it is necessary to extract the feature amount by the operation of the operator on the operation unit. When a region is designated, the feature amount is extracted only from the designated region. Therefore, the processing time can be reduced as compared with the case where the feature amount is extracted from all the regions.
[0121]
According to an eighteenth aspect of the present invention, in the program according to any one of the twelfth to seventeenth aspects, the image comparison function is performed by the feature amount extraction function from the first image and the second image respectively. A feature amount comparison function for comparing the feature amount extracted for each of the plurality of regions for each of the plurality of regions, and matching between the feature amount of the first image and the feature amount of the second image by the feature amount comparison function A match number comparison function of comparing a number with a predetermined match number; and if the match number is determined to be larger than the predetermined match number by the match number comparison function, the first image and the second image are compared. And causing the computer to execute an image specifying function of specifying an image having a high degree of coincidence, so that a feature amount extracted for each of a plurality of regions from the first image and the second image is obtained for each of the plurality of regions. By comparison, high accuracy and simplicity It is possible to perform an image compared to.
[0122]
According to a nineteenth aspect of the present invention, in the program according to the eighteenth aspect, the feature quantity extraction function extracts feature quantities of all areas, and the feature quantity comparison function compares feature quantities of all areas. Therefore, the image comparison can be performed with higher accuracy than when comparing the feature amounts of a partial area.
[0123]
According to a twentieth aspect of the present invention, in the program according to any one of the twelfth to nineteenth aspects, the compressed image is expanded for each of a plurality of regions by a procedure of decoding, inverse quantization, and inverse two-dimensional wavelet transform. By causing the computer to execute the decompression function, the compressed image is decompressed, and as a result, the decompressed image can be displayed on a display device or the like or printed on paper or the like.
[0124]
According to the computer-readable storage medium of the present invention described in claim 21, since the program according to any one of claims 12 to 20 is stored, the program according to any one of claims 12 to 20 is stored. A similar effect is achieved.
[Brief description of the drawings]
FIG. 1 is a functional block diagram for explaining an outline of a JPEG2000 algorithm.
FIG. 2 is a schematic diagram schematically showing an example of each divided component of an original image that is a color image.
FIG. 3 is a schematic diagram schematically showing subbands at each decomposition level when the number of decomposition levels is three.
FIG. 4 is a schematic diagram schematically illustrating an example of a structure of a code stream.
FIG. 5 is a longitudinal sectional view schematically showing the multifunction peripheral according to the first embodiment of the present invention.
FIG. 6 is a block diagram schematically showing an electrical connection of a control system related to image processing in a control system of the multifunction peripheral.
FIG. 7 is a functional block diagram for explaining an outline of image processing in the multifunction peripheral.
FIG. 8 is a flowchart schematically showing a flow of image processing according to the first embodiment of the present invention.
FIG. 9 is an explanatory diagram schematically showing an example of an image divided into a plurality of tile areas.
FIG. 10 is a flowchart schematically illustrating a flow of a feature amount extraction process according to the first embodiment of this invention.
FIG. 11 is a flowchart schematically showing a flow of an image comparison process.
FIG. 12 is an explanatory diagram schematically showing an image comparison process.
FIG. 13 is an explanatory diagram schematically showing a result of an image comparison process.
FIG. 14 is a flowchart schematically illustrating a flow of a feature amount extraction process according to the second embodiment of this invention;
FIG. 15 is an explanatory diagram schematically showing an example of a necessary area.
[Explanation of symbols]
1 Image processing device (composite machine)
7 Reading optical system
17 Printer Engine
31 Storage media (ROM)
33 Operation unit (operation panel)

Claims (21)

Feature amount extracting means for extracting a feature amount for each of the plurality of regions from the first image and the second image, each of which is divided into a plurality of regions and compressed;
An image comparing the first image and the second image based on a feature amount extracted for each of a plurality of regions from the first image and the second image by the feature amount extracting unit; Means of comparison;
An image processing apparatus comprising: The image processing apparatus according to claim 1, wherein the feature amount extracting unit identifies a region attribute of the region and extracts the identified region attribute as a feature amount. The first image and the second image are compressed by a procedure of two-dimensional wavelet transform, quantization, and encoding for each of a plurality of regions,
The image processing apparatus according to claim 2, wherein the feature amount extracting unit identifies a region attribute of the region based on frequency component information included in a two-dimensional wavelet coefficient of the region. The image processing apparatus according to claim 3, wherein the feature amount extraction unit identifies a region attribute of the tile region as a table region attribute based on an HL component and an LH component included in a two-dimensional wavelet coefficient of the region. 2. The image processing apparatus according to claim 1, wherein the feature value extracting unit extracts ruled line information as a feature value based on HL components and LH components included in a two-dimensional wavelet coefficient of the area. An operation unit for receiving an operation by the operator;
An area specifying unit that specifies an area in which a feature amount needs to be extracted based on an operation performed by the operator on the operation unit;
With
The image processing apparatus according to claim 1, wherein the feature amount extracting unit extracts a feature amount of only an area specified by the area specifying unit. The image comparing means includes:
A feature amount comparing unit that compares, for each of a plurality of regions, a feature amount extracted for each of a plurality of regions from the first image and the second image by the feature amount extracting unit;
Coincidence number comparison means for comparing the number of coincidences between the feature quantity of the first image and the feature quantity of the second image by the feature quantity comparison means with a predetermined match number;
An image identification unit that identifies the first image and the second image as images having a high degree of coincidence when the number of coincidences is determined by the coincidence number comparison unit to be greater than a predetermined number of coincidences;
The image processing apparatus according to claim 1, further comprising: The feature amount extraction means extracts feature amounts of all regions,
The image processing apparatus according to claim 7, wherein the feature amount comparing unit compares feature amounts of all the regions. 9. The image processing apparatus according to claim 1, further comprising a reading optical system that optically reads an image from a document. The image processing apparatus according to any one of claims 1 to 9, further comprising decompression means for decompressing the compressed image for each of a plurality of regions by a procedure of decoding, inverse quantization, and inverse two-dimensional wavelet transform. The image processing apparatus according to claim 10, further comprising a printer engine that forms an image on a recording material, the image expanded by the expansion unit. Interpreted by a computer included in the image processing apparatus, the computer,
A feature amount extraction function of extracting a feature amount for each of the plurality of regions from the first image and the second image, each of which is divided into a plurality of regions and compressed;
An image comparing the first image and the second image based on a feature amount extracted for each of a plurality of regions from the first image and the second image by the feature amount extracting function; Compare function,
A program that executes 13. The program according to claim 12, wherein the feature amount extracting function identifies a region attribute of the region and extracts the identified region attribute as a feature amount. The first image and the second image are compressed by a procedure of two-dimensional wavelet transform, quantization, and encoding for each of a plurality of regions,
14. The program according to claim 13, wherein the feature amount extracting function identifies a region attribute of the region based on frequency component information included in a two-dimensional wavelet coefficient of the region. 15. The program according to claim 14, wherein the feature amount extraction function identifies a region attribute of the tile region as a table region attribute based on an HL component and an LH component included in a two-dimensional wavelet coefficient of the region. 13. The program according to claim 12, wherein the feature amount extracting function extracts ruled line information as a feature amount based on HL components and LH components included in a two-dimensional wavelet coefficient of a region. Based on the operation of the operator on the operation unit, causing the computer to execute an area designation function of designating an area in which a feature amount needs to be extracted,
17. The program according to claim 12, wherein the feature amount extracting function extracts a feature amount of only an area specified by the area specifying function. The image comparison function,
A feature amount comparison function for comparing a feature amount extracted for each of a plurality of regions from the first image and the second image by the feature amount extraction function for each of a plurality of regions;
A coincidence number comparison function for comparing the number of coincidences between the characteristic amount of the first image and the characteristic amount of the second image by the characteristic amount comparison function with a predetermined coincidence number;
An image identification function for identifying the first image and the second image as images having a high degree of coincidence when the number of coincidences is determined by the coincidence number comparison function to be greater than a predetermined number of coincidences;
18. The program according to claim 12, which causes the computer to execute the following. The feature amount extraction function extracts feature amounts of all regions,
19. The program according to claim 18, wherein the feature amount comparison function compares feature amounts of all areas. 20. The non-transitory computer-readable storage medium according to claim 12, wherein the program causes the computer to execute a decompression function for decompressing a compressed image for each of a plurality of regions by a procedure of inverse quantization, inverse two-dimensional wavelet transform. A computer-readable storage medium storing the program according to claim 12.

Images