JP2001319231A - Device and method for processing image, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suitably retrieve an image approximate to or coincident with a new image in an image database and to prevent double registration in the image database. SOLUTION: With respect to an image to be registered, an area dividing processing is executed (step S202) and a registered image having the coincident number of divided areas is retrieved. Besides, narrowing based on a distance between images is performed as needed (steps S204 and S205). Finally, when there is the coincident or approximate registered image, any image is not newly registered (step S206).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to image processing, and more particularly, to filing an image and retrieving the filed image.

[0002]

2. Description of the Related Art Conventionally, in an image filing system for storing an image including a document or the like, an image captured by an image scanner is added with an index of a keyword for search and stored, thereby facilitating a search.

[0003]

However, in the conventional filing system, in order to prevent the newly registered image from being registered twice with the already registered and stored registered image, it is necessary to store the newly registered image. The registered images must be searched manually by keyword search or by displaying a list of all registered images, and humans must visually confirm that they match the images to be newly registered. ,
It took time. For this reason, a situation has arisen in which double registration is permitted.

That is, in the conventional filing system, the function of searching for a registered image that approximates or matches a new image is not user-friendly.

Accordingly, an object of the present invention is to provide an image processing apparatus, an image processing method, and a recording medium capable of suitably searching for an image that is close to or coincident with a new image, and preventing double registration. Is to provide.

[0006]

According to the present invention, there is provided an image processing apparatus for storing a given image as a registered image, wherein the image processing apparatus performs a region dividing process on the given image.
Processing means for calculating at least the number of divided areas; search means for searching for the registered image that is similar or identical to the image; and determining means for determining whether to save the image as the registered image. A storage unit for storing the image determined to be stored as a registered image, together with the number of regions obtained by the processing unit, wherein the determination unit approximates the image by the search unit or When the matching registered image is searched, it is determined that the image is not stored as the registered image, and the search unit determines that the number of the areas of the image matches the number of the areas. Is searched as a first search candidate.

Further, according to the present invention, there is provided an image processing apparatus for retrieving an image that is similar or identical to a given image from registered images registered in advance.
A recording unit that records the number of areas obtained by executing the area division processing on the image, and executes the area division processing on the given image to calculate at least the number of divided areas Processing means; and a search means for searching, as a first search candidate, the registered image in which the number of areas obtained by the processing means for the given image matches the number of areas, as a first search candidate. Is provided.

Further, according to the present invention, there is provided an image processing method for storing a given image as a registered image, wherein a region dividing process is performed on the given image and at least the number of divided regions is determined. A calculating step, a search step of searching for the registered image that approximates or matches the image, a determining step of determining whether to save the image as the registered image, and a determination of saving as the registered image And storing the obtained image together with the number of regions obtained in the processing step. In the determining step, in the searching step, the registered image that approximates or matches the image is searched. If it is determined that the image is not stored as the registered image, in the search step, the number of the region of the image and the registered image in which the number of the region matches the registered image Image processing method characterized by retrieving a first search candidates is provided.

Further, according to the present invention, there is provided an image processing method for retrieving an image which is similar or identical to a given image from registered images registered in advance.
A recording step in which the image is recorded together with the number of areas obtained by executing the area division processing; and executing the area division processing on the given image to calculate at least the number of divided areas. A processing step; and a search step of searching, as a first search candidate, the number of areas obtained in the processing step for the given image and the registered image having the same number of areas as a first search candidate. An image processing method is provided.

According to the present invention, in order to store a given image as a registered image, a computer executes a region dividing process on the given image and calculates at least the number of divided regions. Processing means for performing the search, a search means for searching for the registered image that approximates or matches the image, a determination means for determining whether to save the image as the registered image, and the image determined to be stored as the registered image Is a storage medium storing a program functioning as a storage unit for storing the number of regions obtained by the processing unit, wherein the determination unit approximates or matches the image by the search unit. When the registered image is searched, it is determined that the image is not stored as the registered image, and the search unit determines the number of the areas of the image, Recording medium characterized by retrieving the registered image number of frequency matches the first search suggestions are provided.

Further, according to the present invention, in order to search for an image that is similar or identical to a given image from registered images registered in advance, a computer is used to search the registered image for the image. A recording unit that records the number of areas obtained by executing the area division processing, a processing unit that executes the area division processing on a given image, and calculates at least the number of divided areas, A recording medium is provided which records a program functioning as a search unit for searching, as a first search candidate, the registered image having the same number of regions as the number of regions obtained by the processing unit with respect to the image.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.

FIG. 6 is a diagram showing an example of hardware for realizing the image filing system according to one embodiment of the present invention. As shown in FIG. 6, the present system can be realized on a general computer system.

In FIG. 6, reference numeral 600 denotes a bus for transmitting address signals and data, 601 a CPU for control, 602 a ROM for storing a program for booting a BIOS or an OS, and 603 for an OS or various programs. Is a RAM used for loading a file and a work area.

An external storage device 604 stores an image database, stores an OS and various programs, or stores a temporary file of work data.
Reference numeral 5 denotes a display for displaying a document image, various messages, and the like. Reference numeral 606 denotes an image scanner interface. Reference numeral 607 denotes an image scanner that reads a document and converts it into a document image.

FIG. 1 is a flowchart for explaining a new image registration process. Hereinafter, in the present embodiment, for convenience of explanation, an image provided for new registration is referred to as an input image, and an image that has been registered and stored is referred to as a registered image.

The input image is, for example, an image scanner 6
07, a document or the like is read and stored as a color image or a monochrome multi-valued image or a binary image in the external storage device 6.
04 or the like.

The input image can also be obtained by converting document data created by application software into an image in a bitmap format or the like.

FIG. 8 illustrates steps for converting document data created by application software into a bitmap image.

In step S802, the input document data (step S801) is converted into a GDI format 803 by application software such as word processing software (for example, Microsoft Word, Ichitaro: both are registered trademarks) (step S803).

In step S804, the printer driver or facsimile driver converts the image into a bitmap image (step S805).

Next, returning to FIG. 1, in step S101, a region dividing process is performed on the input image. The region dividing process is a process of dividing an image into blocks (regions) according to the type of the content. For example, it is a process of dividing the image into blocks according to types such as a text portion, an image portion, and a table portion. Various specific contents of such a region dividing process have been proposed.
-068301 and the like.

FIG. 7 is a diagram showing an example of the area division processing. In FIG. 7, reference numeral 701 denotes an entire image;
703 and 704 are text blocks, 705 and 706
Indicates an image block. Note that block 70
The bold lines in 3 and 704 represent simplified character strings. In the example of FIG. 7, the types of blocks are divided into two types, that is, text and images. However, it goes without saying that the types of blocks may be subdivided into more types.

The result of area division of the input image is shown in FIG.
It is stored in the RAM 603 as one block information. FIG. 4 is a diagram illustrating this block information.

The block information includes a block information header,
It consists of block information data 1 to block information data n of each of the divided blocks.

The block information header includes, for example, information on the total number of blocks, the number of text blocks, the number of image blocks, and the number of indistinguishable blocks. In addition, each block information data includes, for example, information relating to a block ID, a block type, block coordinate information, a block width, and a block height. In the present embodiment, the coordinates of the center of the block are used as the coordinate information of the block, but coordinates other than the center (for example, the coordinates of the upper left vertex of the block) may be used. The program shown in the lower part of FIG. 4 is an example of a program in which the contents of a block information header and block information data are described in C language.

In the following, each block obtained by dividing the area is represented by
B1, B2, B3,..., Bn. As described above, blocks are classified into text blocks and image blocks, and the center coordinates of the block Bi are represented by CX (Bi) and CY.
(Bi).

Then, the text block has its center coordinate CX (Bi) as the first key and its center coordinate CY (Bi) as the second key.
Sort as key. Assume that the sorted text blocks are TB1, TB2,..., TBm.

Similarly, the image blocks are sorted, and the sorted results are IB1, IB2,..., IBk.
The block information data of FIG.
B1, TB2,..., TBm, image block IB1,
IB2,..., Ibk.

The input image is stored as a registered image in a pair with the block information 11 in an image database stored and configured in the external storage device 604, and the storage position is temporarily stored in the RAM 603. .

A page table shown in FIG. 3 is stored in the image database. In FIG. 3, the page ID is a number that can uniquely determine a registered image, and is, for example, an order number assigned in the order in which the registered images are registered.
In FIG. 3, m and k indicate the number of text blocks and the number of image blocks extracted by dividing each registered image into regions. In the figure, the instance pointer indicates the recording position in the external storage device 604 where the pair of the corresponding registered image and block information is recorded.

In step S102, the page ID, the number m + k of text blocks and image blocks, the number k of image blocks, and the instance pointer of the image registered this time are additionally recorded in the page table. Next, in step S103, this table is sorted using m + k as a first key and k as a second key.

As described above, the input image is stored as a registered image. However, if the input image is the same as the existing registered image, it is necessary to prevent double registration. Further, when the registered image is very similar to the existing registered image, the user may not wish to register the image. Therefore, in the present system, registration processing for preventing double registration is performed according to the flowchart of FIG.

In step S202, a region dividing process is performed on the input image. This process is the same as in FIG. As a result, block information as shown in FIG. 4 is obtained.

In step S203, the total number of blocks m + k = n is calculated from the number of text blocks m and the number of image blocks k of the input image, and the registered image which matches the total number of blocks n is referred to by referring to the page table. Refine as search candidates for. If there is only one first search candidate, this may be the final candidate.

In step S204, the degree of coincidence between each registered image of the first search candidates and the input image is determined.
In the present embodiment, the distance between each registered image and the input image is determined based on the size and position of each block. Here, the distance between the registered image and the input image is obtained as follows.

With respect to the input image, the text blocks sorted by the above-described method are denoted by TB′1, TB′2,..., TB′m ′, and the image blocks are denoted by IB′1, IB′2,. , TB1, TB2,..., TBm, image blocks IB1, IB2,. Furthermore, the width and height of the text block TBi are
W (TBi), H (TBi) and the width and height of the image block IBj are represented as W (IBj) and H (IBj).

The distance D between two images is calculated as follows.

[0039]

(Equation 1)

Here, the term mt of the series is mt = min (m, m ').

[0041]

(Equation 2)

Here, the term number ki of the series is as follows: ki = min (k, k '). Then, a distance D (hereinafter, referred to as a first identification function) is calculated as D = Dt + αDi. Here, α is a constant which is experimentally determined in advance so that the image can be identified best. In general,
Since it is considered that an image block is more accurately obtained than a text block, for example, a value of about 2 may be empirically set. In other words, Di is Dt
It works more effectively for identification.

When the distance D between each registered image relating to the first search candidate and the input image is obtained in this way, the search candidates are narrowed down by selecting some registered images from the smaller distance D. For example, narrowing down is performed by selecting only three from the smaller distance D, or by selecting from the smaller distance D such that the number of first search candidates is reduced to a certain ratio (for example, 1/5). In this way, the search is narrowed down and set as a second search candidate. This set of second search candidates is defined as S2.

At this stage, the registered images having the minimum distance D may be narrowed down to only one, and this may be used as the final candidate.

Next, in step S205, further narrowing down is performed as follows.

An image closest to the input image is selected from the set S2 as follows. In the above search,
For both corresponding blocks, the block position and the block size were compared to calculate the distance. This time, the contents of each corresponding block are compared, and further distance calculation is performed. There are comparisons between text blocks and comparisons between image blocks. Text blocks corresponding to the input image and the registered image are TXTB2 and TXTB. For the correspondence, the X coordinate of the center of the block is set to the first key,
This is performed by associating blocks having the same rank when sorting using the Y coordinate as the second key. A character string can be obtained by binarizing these text blocks and performing OCR (optical character recognition). And the character string of TXTB2 and TXT
DP matching of the character string of B (Dynamic Programming:
By dynamic programming), the number of characters in TXTB2 but not in TXTB: n1 The number of characters in TXTB but not in TXTB2: n2 The character strings corresponding to TXTB2 and TXTB differ. The number of characters present: n32, n3 (n32 is the number of characters in TXTB2, n3 is the number of characters in TXTB). DP matching is a known technique that is also disclosed, for example, on page 107 of the information science course “Speech Recognition” (written by Yasunaga Niimi, Kyoritsu Shuppan).

FIG. 5 is an explanatory diagram of an example in which each character string of TXTB2 and TXTB is DP-matched. In the figure, each character string is represented by a thick line. The part E is the part where the character strings match, and the part X is the above 1 (TXTB2, TXT
(The character not in B) and Y are the above 2 (TXTB2
, And the Z portion is the above 3 (a character in which the character string corresponding to TXTB2 and TXTB is different).

As a result, two text blocks TXT
The distance between B2 and TXTB can be calculated as follows.

D (TXTB2, TXTB) = (n1 + n
2 + n32 + n3) / NC Here, NC is the sum of the number of characters of TXTB2 and the number of characters of TXTB.

Thus, the distance is obtained for the corresponding text block. It is also conceivable that the number of text blocks does not match between the input image and the registered image. For example, if the input image has a larger number of text blocks and there is no text block of the registered image corresponding to TXTB2, the distance is 1. The sum of the distances obtained for all the text blocks in this way is called a text block distance.

This time, comparison is made between the input image and the registered image for the image block. Image block IMGB2 of the input image and image block IMGB of the registered image
Shall correspond. The association is performed by associating blocks having the same rank when the X coordinate of the center of the block is sorted as the first key and the Y coordinate as the second key.
For an image obtained by binarizing IMGB2, the ratio of the number of black pixels to the total number of pixels (that is, the number of black pixels ÷ the number of all pixels) ratio (IMGB2) is determined. Similarly, ratio
(IMGB) is obtained.

| Ratio (IMGB2) −ratio (IMGB) | is the distance between IMGB2 and IMGB. If there is no corresponding image block, the distance is set to the maximum value of 1.
Then, the total value of the distances for all the image blocks is referred to as an image block distance.

Now, the detailed identification distance d between the input image and the registered image is obtained by d = text block distance + β × image block distance. Here, β is the same as α described above, and is a weighting factor applied to the image block distance. It is desirable to experimentally determine β so that the image can be identified well, but since the image block can be extracted with higher accuracy (that is, higher reliability) than the text block, it is roughly a value larger than 1 (for example, 2). It may be. The above d is referred to as a second identification function.

Now, the detailed identification distance d between all the registered images of the set S2 and the input image is obtained, and in step S206, the minimum detailed identification distance d0 is compared with a predetermined value δ. If d0 is smaller than δ, it is determined that the registered image of the set S2 giving d0 is a registered image that matches the input image.

Here, δ is a value obtained by a preliminary experiment. For example, a detailed distance (1) between an image of 1000 images that can be read many times by an image scanner under different conditions and an image read under certain conditions
000), and the maximum value of the 1,000 numerical values is defined as δ.

If there is a match in step S206, the input image is not registered in the image database. If there is no match, the process proceeds to step S207.

Step S207 is performed in step S1 described with reference to FIG.
This is exactly the same as the document registration process from 02 to S103.

Although the preferred embodiment of the present invention has been described above, the first and second discriminant functions described above calculate the distance between the registered image and the input image. It may be calculated. For example, it is obvious that the reciprocal of the distance is calculated, that is, the similarity is obtained. When the similarity is used as the identification function, the selection of candidates is narrowed down in descending order of similarity.

In the above embodiment, double registration of images is avoided. However, the same processing can be applied to other uses, for example, an image search device. When performing an image search, a matching registered image can be found from the database by performing the processing of steps S202 to S206 on the input image. Thereafter, by performing a process of extracting the matching registered image, Image search becomes possible.

For example, when a document image is to be searched, there is a document image which is almost the same as the document image to be found at hand, but is slightly different, and there is a use for extracting the original document image from the database. in this case,
If the database is configured as in the above embodiment, a search for a document image most similar to a document image at hand can be performed without manual work. In such an application, since the document image at hand is the result of copying the original multiple times, if the printing condition has deteriorated, a good-printed document must be extracted from the original again. There are cases where you want to.

An object of the present invention is to supply a storage medium (or a recording medium) in which program codes of software for realizing the functions of the above-described embodiments are recorded to a system or an apparatus, and to provide a computer (a computer) of the system or the apparatus. It is needless to say that the present invention can also be achieved by a CPU or an MPU) reading and executing the program code stored in the storage medium. In this case, the program code itself read from the storage medium implements the functions of the above-described embodiment, and the storage medium storing the program code constitutes the present invention. Also,
When the computer executes the readout program code, not only the functions of the above-described embodiments are realized, but also the operating system (OS) running on the computer based on the instructions of the program code.
It goes without saying that a case where the functions of the above-described embodiments are implemented by performing some or all of the actual processing, and the processing performs the functions of the above-described embodiments.

Further, after the program code read from the storage medium is written in the memory provided in the function expansion card inserted into the computer or the function expansion unit connected to the computer, the program code is read based on the instruction of the program code. Needless to say, the CPU included in the function expansion card or the function expansion unit performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

[0063]

As described above, according to the present invention, it is possible to suitably search for an image that is similar to or coincides with a new image, and to prevent double registration.

[Brief description of the drawings]

FIG. 1 is a flowchart illustrating a registration process of a new image.

FIG. 2 is a flowchart illustrating a registration process of a new image while preventing double registration.

FIG. 3 is an explanatory diagram of a page table.

FIG. 4 is an explanatory diagram of block information.

FIG. 5 is an explanatory diagram of DP matching of a text block.

FIG. 6 is a diagram illustrating an example of hardware for realizing an image filing system according to an embodiment of the present invention.

FIG. 7 is a diagram illustrating an example in which a region division process is performed.

FIG. 8 is a flowchart of a process for obtaining an input image from document data.

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06K 9/20 340 G06K 9/20 340C G06T 1/00 200 G06T 1/00 200E 200D 7/40 100 7 / 40 100A 100C 7/60 110 7/60 110 F-term (reference) 5B029 AA01 CC29 5B050 AA09 BA10 BA15 BA16 BA20 CA07 EA06 EA18 FA02 GA08 5B075 ND06 NR03 NR14 PR06 QM08 5L096 AA06 BA20 DA02 EA43 FA44 FA03 FA53 FA53 FA53 FA53 KA09 KA15

Claims (30)

[Claims] 1. An image processing apparatus for storing a given image as a registered image, performing a region division process on the given image, and calculating at least the number of divided regions; A search unit that searches for the registered image that is similar to or coincides with the image; a determination unit that determines whether to save the image as the registered image; and the image that is determined to be stored as the registered image, Storage means for storing together with the number of regions obtained by the processing means, wherein the determination means, when the registered image that is close to or coincident with the image is searched by the search means, It is determined that the image is not stored as the registered image, and the search means searches for the registered image in which the number of the areas of the image matches the number of the areas as a first search candidate. An image processing apparatus, comprising: 2. The processing unit also calculates a position and a size of each of the divided areas, and the storage unit obtains the image determined to be stored as a registered image by the processing unit. When the plurality of first search candidates are present, the search means further stores the position and size of each of the regions in the image among the first search candidates. The image processing apparatus according to claim 1, wherein the registered image in which the position and the size of the area are similar is searched as a second search candidate. 3. A character recognition process is performed on a text area of the area of the registered image related to the image and the second search candidate, and the registration related to the image and the second search candidate is performed. Means for calculating the degree of coincidence of the text area corresponding to an image, wherein the search means has a high degree of coincidence of the calculated text area when there are a plurality of second search candidates. The image processing apparatus according to claim 2, wherein is searched for. 4. An image feature extracting unit extracts feature data from an image region among the regions of the registered image relating to the image and the second search candidate, and extracts the image and the second search candidate. Means for calculating the degree of coincidence of the corresponding image area with the registered image according to the above, wherein the search means, when there are a plurality of the second search candidates, matches the calculated image area The image processing apparatus according to claim 2, wherein a search is made for a high degree. 5. The image feature extraction means executes a binarization process on an image area, calculates a ratio of the number of black pixels to the total number of pixels, and uses the ratio of the number of black pixels as feature data. The image processing apparatus according to claim 4, wherein: 6. An image processing apparatus for searching for an image that is similar or identical to a given image from registered images registered in advance, and executes a region division process on the registered image. Recording means for recording together with the number of areas obtained by performing the above, processing means for executing area division processing on a given image and calculating at least the number of divided areas, The registered image in which the number of areas obtained by the processing means matches the number of areas is first registered.
An image processing apparatus, comprising: a search unit that searches as a search candidate of (i). 7. The recording unit records the registered image together with the number, position, and size of an area obtained by executing an area dividing process on the image. In the case where there are a plurality of one search candidates, further, among the first search candidates, the registered image in which the position and size of each area of the image are similar to the position and size of the area is set to the second search candidate. The image processing apparatus according to claim 6, wherein a search is performed as a search candidate. 8. A character recognition process is performed on a text region of the region of the registered image related to the image and the second search candidate, and the registration related to the image and the second search candidate is performed. Means for calculating the degree of coincidence of the text area corresponding to an image, wherein the search means has a high degree of coincidence of the calculated text area when there are a plurality of second search candidates. The image processing apparatus according to claim 7, wherein is searched. 9. An image feature extracting unit extracts feature data from an image area of the registered image related to the image and the second search candidate, and extracts the image and the second search candidate. Means for calculating the degree of coincidence of the corresponding image area with the registered image according to the above, wherein the search means, when there are a plurality of the second search candidates, matches the calculated image area The image processing apparatus according to claim 7, wherein a search is performed for a high degree. 10. The image feature extraction means executes a binarization process on an image area, calculates a ratio of the number of black pixels to the total number of pixels, and uses the ratio of the number of black pixels as feature data. The image processing apparatus according to claim 9, wherein: 11. An image processing method for storing a given image as a registered image, comprising: performing a region dividing process on the given image to calculate at least the number of divided regions; A search step of searching for the registered image that is similar to or matching the image; a determining step of determining whether to save the image as the registered image; and the image determined to be saved as the registered image, And a saving step of saving together with the number of regions obtained in the processing step, wherein in the determining step, in the search step, if the registered image that is close to or matching the image is searched, It is determined that the image is not stored as the registered image. In the search step, the registered image having the same number of regions as the number of regions in the image is defined as a first search candidate. An image processing method characterized by performing a search. 12. In the processing step, a position and a size of each of the divided areas are also calculated. In the storing step, the image determined to be stored as a registered image is obtained in the processing step. In the search step, when there are a plurality of the first search candidates, the position and the size of each area of the image in the first search candidates are further stored in the search step. 12. The image processing method according to claim 11, wherein the registered image in which the position and the size of the area are similar is searched as a second search candidate. 13. A character recognition process is performed on a text region of the region of the registered image related to the image and the second search candidate, and the registration related to the image and the second search candidate is performed. Calculating a degree of coincidence of the text area corresponding to an image, wherein in the search step, when there is a plurality of second search candidates, the degree of coincidence of the calculated text area is high The image processing method according to claim 12, wherein is searched. 14. An image feature extracting step of extracting image data from an image area of the area of the registered image related to the image and the second search candidate, wherein the image and the second search candidate are extracted. Calculating the degree of coincidence of the corresponding image area with the registered image according to the above. In the search step, when there is a plurality of the second search candidates, the calculated image area coincidence 13. The image processing method according to claim 12, wherein a search is made for a high degree. 15. In the image feature extraction step, a binarization process is performed on the image area, a ratio of the number of black pixels to the total number of pixels is calculated, and the ratio of the number of black pixels is used as feature data. 15. The image processing method according to claim 14, wherein: 16. An image processing method for retrieving an image that approximates or matches a given image from registered images registered in advance, wherein the registered image is subjected to a region division process on the image. A recording process of recording the number of areas obtained by performing the above-mentioned process, a process of executing a region dividing process on a given image, and calculating at least the number of divided regions; An image processing method comprising: a search step of searching, as a first search candidate, the number of areas obtained in the processing step and the registered image having the same number of areas as a first search candidate. 17. In the recording step, the registered image is
The number, the position, and the size of the area obtained by executing the area division processing on the image are recorded. In the search step, when there are a plurality of the first search candidates, the first 17. The search image of claim 16, wherein the registered image in which the position and size of each area of the image and the position and size of the area are similar is searched as a second search candidate. Image processing method. 18. A character recognition process is performed on a text area in the area of the registered image related to the image and the second search candidate, and the registration related to the image and the second search candidate is performed. Calculating a degree of coincidence of the text area corresponding to an image, wherein in the search step, when there is a plurality of second search candidates, the degree of coincidence of the calculated text area is high 18. The image processing method according to claim 17, wherein is searched. 19. An image area of the registered image related to the image and the second search candidate, wherein feature data is extracted by an image feature extraction step, and the image and the second search candidate are extracted. Calculating the degree of coincidence of the corresponding image area with the registered image according to the above. In the search step, when there is a plurality of the second search candidates, the calculated image area coincidence 18. The image processing method according to claim 17, wherein a search is performed for a high degree. 20. In the image feature extraction step, a binarization process is performed on the image area, a ratio of the number of black pixels to the total number of pixels is calculated, and the ratio of the number of black pixels is used as feature data. 20. The image processing method according to claim 19, wherein: 21. A computer for executing a region dividing process on the given image to calculate at least the number of divided regions, in order to store the given image as a registered image. Search means for searching for the registered image that is similar to or matched with, determining means for determining whether to save the image as the registered image, and processing the image determined to be stored as the registered image by the processing means. A storage medium storing a program functioning as a storage unit for storing the number of obtained regions together with the number of obtained regions, wherein the determination unit searches for the registered image that is similar to or matches the image by the search unit. If the number of areas is equal to the number of areas in the image, the search unit determines that the image is not stored as the registered image. A storage medium for searching for the registered image as a first search candidate. 22. The processing means also calculates the position and size of each of the divided areas, and the storage means obtains the image determined to be stored as a registered image by the processing means. When the plurality of first search candidates are present, the search means further stores the position and size of each of the regions in the image among the first search candidates. 22. The recording medium according to claim 21, wherein the registered image in which the position and size of the area are similar is searched as a second search candidate. 23. The program, comprising: executing a character recognition process on a text area of the area of the registered image relating to the image and the second search candidate; Means for calculating the degree of coincidence of the text area corresponding to the registered image relating to the search candidate of the search candidate, the search means, when there is a plurality of the second search candidates, 23. The recording medium according to claim 22, wherein a search is made for a text area having a high degree of coincidence. 24. The program according to claim 24, wherein the program causes a computer to extract feature data for an image area of the area of the registered image related to the image and the second search candidate. Means for calculating the degree of coincidence of the corresponding image area with the registered image relating to the second search candidate, the search means comprising a plurality of the second search candidates 23. The recording medium according to claim 22, wherein in that case, a search is made for a high degree of coincidence between the calculated image areas. 25. The image feature extracting means executes a binarization process on an image area, calculates a ratio of the number of black pixels to the total number of pixels, and uses the ratio of the number of black pixels as feature data. The recording medium according to claim 24, wherein: 26. A computer executes a region division process on the registered image in order to search for an image that is similar or identical to a given image from registered images registered in advance. Recording means for recording together with the number of areas obtained by the above, processing means for performing area division processing on a given image, and calculating at least the number of divided areas, and said processing means for the given image The registered image in which the number of regions obtained by
A recording medium on which a program functioning as a search means for searching as a search candidate is recorded. 27. The recording unit records the registered image together with the number, position, and size of regions obtained by executing region division processing on the image. In the case where there are a plurality of one search candidates, further, among the first search candidates, the registered image in which the position and size of each area of the image are similar to the position and size of the area is set to the second search candidate. The recording medium according to claim 26, wherein a search is performed as a search candidate. 28. The program, wherein the computer executes a character recognition process on a text area of the image and the area of the registered image related to the second search candidate, and executes the character recognition processing on the image and the second search candidate. A program that functions as a unit that calculates the degree of coincidence of the corresponding text area with the registered image related to the search candidate, wherein the search unit includes a plurality of second search candidates. 28. The recording medium according to claim 27, wherein a search is made for a text area having a high degree of coincidence. 29. The program according to claim 29, wherein the program causes the computer to: extract image feature data for an image area of the image and the area of the registered image relating to the second search candidate; Means for calculating the degree of coincidence of the corresponding image area with the registered image relating to the second search candidate, the search means comprising a plurality of the second search candidates 28. The recording medium according to claim 27, wherein, in such a case, a search is made for one having a high degree of coincidence between the calculated image areas. 30. The image feature extraction means executes a binarization process on an image area, calculates a ratio of the number of black pixels to the total number of pixels, and uses the ratio of the number of black pixels as feature data. 30. The recording medium according to claim 29, wherein: