JPH11203410A - Method and device for processing image and storage medium therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make compatible a recognition rate and operability in the case of recognizing a document mixing different languages by extracting an area from an input image for each attribute, discriminating language types for each area and recognizing characters corresponding to the discriminated language type. SOLUTION: By an image input part 2, an original image is inputted. This image is inputted from another terminal or storage medium through a scanner or communication I/F and stored in an MEM. An area identifying part 4 identifies the inputted original image. By a Japanese/English discriminating part 6, it is discriminated whether the area is in Japanese or English. By a Japanese/English discrimination control part 8, the Japanese/English discriminating part is controlled. Corresponding to whether or not automatic language discrimination is to be performed by an input means and the input of the instruction of a recognition mode in the case of not performing the automatic discrimination, a control signal expressing the instruction is generated. A character recognizing part 10 recognizes characters in Japanese or English. These processes are executed by a CPU according to a control program stored in the MEM.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method and apparatus and a storage medium thereof, and more particularly to a process for recognizing a character in an image of a document in which different languages are mixed.

[0002] The present invention relates to an image processing method and apparatus and a storage medium thereof, and more particularly to a process of analyzing an image which can be divided into a plurality of blocks.

[0003]

2. Description of the Related Art Conventional character recognition processing is basically performed for a single language. For example, a Japanese language character recognition device is used for Japanese and a English language character recognition device is used for English. Had gone. However, in the case of a Japanese-language specialized character recognition device, since an alphabet is often included as a target character, it is possible to recognize a language composed of the alphabet, for example, English. However, in such a case, it is undeniable that the recognition accuracy of the English part is deteriorated as compared with the English specialized character recognition device.

In order to eliminate such a drawback, the user visually determines whether each sentence is English or Japanese, and instructs to specify a Japanese sentence portion and perform character recognition specialized in Japanese. There is also a character recognizing device for designating an English sentence portion to perform character recognition specialized in English.

[0005]

However, in the character recognition apparatus described in the above-mentioned conventional example, when recognizing a document in which different languages are mixed, it is difficult to perform character recognition in Japanese for both Japanese and English sentences. The recognition rate and usability are not compatible, as the specification of the sentence part and the instruction of the recognition language type must be repeated twice in Japanese and English.
I had to take one or the other. This has also been a major obstacle to the spread of character recognition devices.

[0006]

In order to solve the above problems, the present invention extracts a region for each attribute from an input image, determines a language type for each region, and converts the image of each region into the image. Provided are an image processing method and apparatus for recognizing characters according to a determined language type, and a storage medium therefor.

In order to solve the above-mentioned problem, the present invention preferably determines the language type by analyzing image characteristics of the extracted area.

In order to solve the above-mentioned problem, the present invention preferably determines the language type by identifying an attribute of the language type set for the extracted area.

In order to solve the above problems, the present invention preferably comprises a plurality of character recognition methods corresponding to different languages, and uses a character recognition method suitable for the language type according to the determined language type. Perform recognition.

[0010] In order to solve the above-mentioned problem, the present invention preferably determines whether or not an instruction for automatic language type determination has been made, and if it is determined that the instruction for automatic language type determination has been issued, The language characteristics are determined by analyzing the image characteristics of the extracted area.

In order to solve the above-mentioned problem, the present invention preferably determines the language type by weighting a specific language type with a predetermined weight.

In order to solve the above problem, the present invention preferably controls the determination of the language type so that a language type having many target characters is preferentially selected.

In order to solve the above-mentioned problem, the present invention preferably determines the language type selectively in a region having a sentence attribute.

In order to solve the above-mentioned problem, the present invention preferably inputs a language type instruction from a user, and recognizes the input image information in accordance with the input language type.

In order to solve the above problem, the present invention preferably stores the result of the character recognition by adding document identification information.

In order to solve the above-mentioned problem, the present invention preferably stores the character recognition result and the input image information together with document identification information.

In order to solve the above problem, the present invention preferably stores the result of the character recognition and information obtained by compressing the input image information, with document identification information added thereto.

In order to solve the above problem, the present invention preferably stores the result of the character recognition and the layout information of the extracted area with document identification information added thereto.

In order to solve the above problem, the present invention preferably outputs a result of character recognition in which the document identification information is added and stored in accordance with the specification of the document identification information.

In order to solve the above-mentioned problem, the present invention preferably outputs, in accordance with the identification of the document identification information, a character recognition result and image information stored with the document identification information added.

In order to solve the above-mentioned problem, the present invention preferably provides, in accordance with the identification of document identification information, a character recognition result and image information compressed information which are stored with the document identification information added. Output.

In order to solve the above-mentioned problem, the present invention preferably outputs, in accordance with the identification of the document identification information, a character recognition result added with the document identification information and stored, and layout information of an area. .

[0023]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 5 is a block diagram of the apparatus according to the present invention.

Reference numeral 51 denotes a CPU (central processing unit).
The processing according to the present invention is controlled in accordance with the control program stored in the MEM 53. The processing shown in the flowchart described later is also executed under the control of the CPU 51. 53
Is a MEM (composed of RAM and ROM) and CP
A control program for processing executed by U51, various parameters used for the processing, an input image, a dictionary for character recognition, etc.
Various data is stored in the MEM 53. Reference numeral 54 denotes a display such as a CRT or an LCD, which displays an input image, text of a processing result, an operation instruction screen, a recognition result read from a file in accordance with the document identification information specified by the input unit 59, and the like. The character recognition result stored in the MEM 53 is displayed on the display unit 54, and editing such as selecting a correct character from a plurality of candidate characters using the input unit 59 is possible. Reference numeral 55 denotes a printer such as an LBP or BJ printer, which prints images, texts, and the like. Reference numeral 56 denotes a scanner, which optically reads an image of a document and inputs it to the apparatus as an electrical signal. A communication I / F 57 controls transmission and reception of data via a public line, a LAN, or the like. The input of an image according to the present invention and the output of the result of processing the image can be transmitted and received to and from another terminal via the communication I / F. Reference numeral 58 denotes a storage medium that is detachable from the apparatus and can be read and further written by a computer, such as a CD-ROM, a CR-R, and an FD. The image according to the present invention may be read from the storage medium 58 and the processing result may be written to the storage medium 58. MEM5
3 may be installed from another terminal via the communication I / F, or may be installed from the storage medium 58. Reference numeral 59 denotes input means such as a keyboard and a pointing device. Instructions from the operator are made through the input means 59. A bus 50 transmits and receives data between the respective units.

FIG. 1 is a schematic view of the process of the apparatus according to the present invention. In the present embodiment, a case where Japanese and English are to be recognized will be described as an example.

In FIG. 1, reference numeral 2 denotes an image input unit for inputting an original image, which is input from another terminal via a scanner 56 or a communication I / F 57 or from a storage medium 58, and
3 is stored. 4 is an area identification part of the input original image, 6
Is a Japanese-English discriminator that determines whether the area is Japanese or English, 8
Is a Japanese-English discriminating control unit that controls the Japanese-English discriminating unit.
A control signal representing the instruction is generated. Reference numeral 10 denotes a character recognizing unit that performs Japanese and English character recognition, and each of these processes is performed as shown in a flowchart described later.
3 is executed by the CPU 51 according to the control program stored in the CPU 3.

Next, the operation will be described.

The original image obtained from the image input unit 2 is
As shown in FIG. 2, the area identification unit 4 divides each attribute into small areas (hereinafter, referred to as blocks), and identifies a block number, a block attribute, a block size, and a block position. The area identification process performed by the area identification unit 4 analyzes the arrangement of black pixels in the input document image, determines attributes such as text, figures, images, and separators, and sets a group of images having the same attribute. As one block. Further, even if the texts are the same, the line directions are different or the line spacing is wider than a specified value, it is determined that the text is in a different column, and they are identified as different blocks.

When a block is identified, a block number is automatically assigned to each block, for example, in order from the top, and width information and height information indicating the size of the block, and the position of the block are defined, for example, MEM5 is set as block data with the coordinate data at the upper left corner as a set.
3 is stored. By referring to the block data, an image of a desired one block of the input image can be extracted, and the arrangement of texts, figures, images, separators, and the like in the document can be reproduced. On the other hand, in the Japanese-English discrimination control unit 8, the user sets whether or not to perform automatic Japanese-English discrimination, and the recognition language mode set when the user does not set the automatic Japanese-English discrimination. A control signal corresponding to Japanese recognition or English recognition) is generated and input to the Japanese-English discrimination unit 6. The Japanese-English discriminating section 6 judges whether the character area is Japanese or English based on the control signal generated by the Japanese-English discriminating control section 8 and the area attribute set by the area identifying section 4.

The image data of the character area to which the Japanese or English attribute is added is sent to the character recognizing unit 10, and the character is recognized according to the language specified by each attribute.
The character recognition unit 10 performs a Japanese character recognition routine in S406 and an English character recognition routine in S408, which will be described later. Each of these may be provided with a separate character recognition unit, or a character image cutout. Alternatively, a recognition dictionary may be provided for each language, and a matching algorithm may be used in common.

When the character recognizing process in the character recognizing unit 10 is completed, the result of the character recognition is stored (filed) with document identification information added thereto. This document identification information may be input by a user's keyboard operation, or may be extracted from a character recognition result. Alternatively, the input date and time, the time, the user ID, and the like may be automatically given. What is necessary is that each document is different so that it can be used as an index when reading from a file later. The data to be filed may include not only the character recognition result but also the input image and the layout information of the data area obtained by compressing the input image.

Here, the Japanese-English discriminating section 6 and the character recognizing section 10 will be described in detail.

FIG. 3 is a flowchart showing the flow of the process of the Japanese-English discriminating section 6.

S302 is a routine for reading a control signal generated in the Japanese-English discrimination control unit 8 in accordance with a user's instruction and stored in the MEM 53. When there are a plurality of document images to be processed, the control signal stored in the MEM 53 is selectively read in association with the image currently specified as the processing target. This control signal indicates whether or not to perform automatic Japanese-English discrimination, and indicates a recognition language mode when no discrimination is performed.

In S304, it is determined based on the control signal read in S302 whether or not to perform automatic determination between Japanese and English. If so, control is passed to S306, otherwise to S322.

S306 is a routine for reading the attribute of each block divided as shown in FIG. 2 by the area identifying unit 4, and reads the block data stored in the MEM 53 one by one in the order of the block numbers.

In step S308, it is determined based on the block attribute in the block data read in step S306 whether or not the block is a text area and the process should proceed to step S310. Control is transferred to Japanese-English discrimination processing for. If it is determined in S308 that the block is not a text area and the process should not proceed to S310, control is returned to S306, and the next block data is stored in M
Read from EM53. In the example of FIG. 2, since the text attributes are given to the blocks 2, 3, and 4, these blocks are determined to be a sentence area, and become the target blocks of the Japanese-English determination processing of S310 and subsequent steps.

S310 is a routine for analyzing the image characteristics of the block to determine whether the language type is Japanese or English, based on the block position and size data in the block data read in S306. ME that can be specified
This is a process performed on the image of the corresponding area from the input image of M53. This is, for example, JP-A-8-339424, JP-A-8-30
This can be achieved by applying a technology for automatically discriminating language types such as Japanese and English from the features of the input image described in 5792.

In S312, a new attribute, that is, a language type is set for the block data based on the result determined in S310 and stored in the MEM 53. The setting here is
Japanese, English and unknown. “Unknown” means that S31
This is set when the probability of the Japanese language and the probability of the English language in the Japanese-English discrimination routine at 0 are lower than the threshold.

In S314, S306 to S31 are executed for all blocks of the original image stored in the MEM 53.
It is determined whether or not the language type determination process 2 has been performed. If all the processes have been completed, the process proceeds to S316.
The control is shifted to S306.

In steps S316, S318, and S320, the language type of all blocks of the original is checked. If at least one Japanese block exists, the unknown block is regarded as a Japanese block. If all of the blocks are English blocks, the routine is a routine for determining the language type of an unknown block to be set as an English block. In this routine, S31
In step 6, the attribute of the language type of all blocks stored in the MEM 53 is read, and it is determined whether there is at least one block in which the Japanese attribute is set. If it is determined that there is a block, the reading is performed in step S316. The language type of the block whose attribute is unknown is set to Japanese and stored in the MEM 53 (S31).
8) If it is determined in S316 that there is no Japanese block, the language type is set to English for the block whose attribute of the block read in S316 is unknown and stored in the MEM 53. (S320). S3
The determination criteria of 16 may be determined in advance, and a predetermined number (1, 2,...) Of a specific language type block or a predetermined ratio (5%, 10%,...) Is also good.

This is because even if the unknown block is actually an English block but the Japanese block is set in S318, the Japanese character recognition performed by the character recognition unit 10 includes the recognition of the alphabet (ie, the recognition of the alphabet). ,
In addition to Japanese kana, kanji, numbers, symbols, etc., the alphabet is also used as a recognition target character), so it does not fall into the inability to recognize, but in the opposite case, that is, when the Japanese block is incorrectly set as the English block Is unaware that it is stricter to make such a judgment. That is, it is easier to select a language type having a large number of characters to be recognized. More preferably, as described in the present embodiment, when the recognition target character of one language type includes the recognition target character of the other language type, it is easy to select the language type containing the recognition target character. Good.

S322 is a routine for setting a recognition language mode when it is determined in S304 that automatic determination of Japanese and English is not performed. This setting is performed by the operator via the input means 59 and based on the control signal generated by the Japanese-English discrimination control unit 8 in accordance with the language type specified.

In S324, if it is determined in S322 that the mode is the Japanese recognition mode, all text areas in the original are set to Japanese,
Routine to be stored in MEM53.

In S326, if it is determined in S322 that the mode is the English recognition mode, all the text areas in the document are set to English, and ME
Routine stored in M53.

FIG. 4 is a flowchart showing the flow of the character recognition process performed by the character recognition unit 10.

S402 is a routine for reading, from the block data stored in the MEM 53, block data of a text block whose attribute is set as text and image data in the block.

In step S404, a language type attribute of the block is determined based on the attribute of the block data read in step S402. The attributes determined by this routine are S310,
These are the attributes of the language types set in S318, S320, S324, and S326.

In S406, if it is determined in S404 that the language is Japanese,
Routine for performing Japanese character recognition in the character recognition unit 10.

S408 is a routine for recognizing English characters in the character recognizing section 10 when it is determined in S404 that the character is English.

In step S410, it is determined whether or not the character recognition processing in steps S402 to S408 has been performed on all the text blocks of the original document stored in the MEM 53. If the processing blocks still remain, the control proceeds to step S402. And reads the next sentence block.

As described above, according to the present invention,
It provides a user-friendly interface for its control based on Japanese-English automatic discrimination technology, and can also construct a unique post-processing character recognition device for the processing of unknown blocks. Even in a mixed document, the recognition accuracy and the usability can be improved at the same time, which greatly contributes to the spread of the character recognition device.

In the present embodiment, the description has been given by taking as an example the case of determining the language type between Japanese and English. In this case, basically, it is possible to cope only by replacing the S310 Japanese / English discriminating unit in FIG. 3 with a language type discriminating routine suitable for it.

[0055]

As described above, according to the present invention,
It provides a user-friendly interface for its control, based on the automatic language type discrimination technology, and can build a character recognition device that takes into account post-processing at the time of erroneous determination. In addition, it is possible to simultaneously improve the usability and greatly contribute to the spread of the character recognition function.

As described above, according to the present invention, a sentence divided into a plurality of regions included in a manuscript is subjected to character recognition according to the language type of each region by determining the language type for each region. A document in which a plurality of languages are mixed can be recognized with high accuracy and high speed.

[Brief description of the drawings]

FIG. 1 is a process schematic diagram of a character recognition device according to the present invention.

FIG. 2 is a view showing an example of an area identification result;

FIG. 3 is a flowchart showing the flow of processing of a Japanese-English discriminator 6;

FIG. 4 is a flowchart showing a flow of processing of a character recognition unit 10;

FIG. 5 is a configuration diagram of an apparatus according to the present invention.

Claims (35)

[Claims] 1. An area is extracted for each attribute from an input image, a language type is determined for each area, and characters in the image of each area are recognized according to the determined language type. Image processing method. 2. The image processing method according to claim 1, wherein the language type is determined by analyzing image characteristics of the extracted area. 3. The image processing method according to claim 1, wherein the determination of the language type is performed by identifying an attribute of the language type set for the extracted area. 4. The method according to claim 1, further comprising a plurality of character recognition methods corresponding to different languages, wherein character recognition is performed by a character recognition method suitable for the language type according to the determined language type. The image processing method described in the above. And determining whether or not an instruction for automatic language type determination has been given. If the determination indicates that an instruction for automatic language type determination has been issued, the image characteristics of the extracted area are analyzed to determine the language. 2. The image processing method according to claim 1, wherein a species is determined. 6. The method according to claim 1, wherein the language type is determined by giving a predetermined weight to a specific language type.
The image processing method according to 1. 7. The image processing method according to claim 1, wherein the determination of the language type is controlled so that a language type having many target characters is preferentially selected. 8. The method according to claim 1, wherein the determination of the language type is selectively performed on a region having a sentence attribute.
The image processing method according to 1. 9. The image processing method according to claim 1, wherein a language type instruction is input by a user, and characters of the input image information are recognized in accordance with the input language type. 10. The image processing method according to claim 1, wherein document identification information is added to a result of the character recognition and stored. 11. The image processing method according to claim 1, wherein a result of the character recognition and the input image information are stored with document identification information added thereto. 12. The image processing method according to claim 1, wherein a result of the character recognition and information obtained by compressing the input image information are stored with document identification information added thereto. 13. The image processing method according to claim 1, wherein the result of the character recognition and the layout information of the extracted area are stored with document identification information added thereto. 14. The image processing method according to claim 10, wherein, in accordance with the specification of the document identification information, a character recognition result to which the document identification information is added and stored is output. 15. The apparatus according to claim 11, wherein, in accordance with the specification of the document identification information, a character recognition result and image information stored with the document identification information added thereto are output.
The image processing method according to 1. 16. The apparatus according to claim 12, wherein, in accordance with the specification of the document identification information, a result of character recognition stored with the document identification information added and information obtained by compressing image information are output. Image processing method. 17. The image according to claim 13, wherein, in accordance with the specification of the document identification information, the result of character recognition and the layout information of the area stored with the document identification information added are output. Processing method. 18. An area extracting unit for extracting an area for each attribute from an input image, a language type determining unit for determining a language type for each area, and an image of each area according to the determined language type. An image processing apparatus comprising: 19. The image processing apparatus according to claim 18, wherein the language type determining unit determines a language type by analyzing image characteristics of the extracted area. 20. The image processing apparatus according to claim 18, wherein the language type determining unit determines by identifying a language type attribute set for the extracted area. 21. The character recognizing unit includes a plurality of character recognizing units corresponding to different languages, and performs character recognition by a character recognizing unit suitable for the language type according to the language type determined by the language type determining unit. 19. The method according to claim 18, wherein
An image processing apparatus according to claim 1. 22. An automatic language type determination control means for switching whether or not to determine the language type by analyzing the image characteristics of the extracted area in accordance with the presence or absence of an instruction for automatic language type determination. 19. The image processing device according to claim 18, wherein 23. The image processing apparatus according to claim 18, wherein the language type determination unit determines a language type by giving a predetermined weight to a specific language type. 24. The image processing apparatus according to claim 18, wherein the language type determination unit controls to preferentially select a language type having many target characters. 25. The image processing apparatus according to claim 18, wherein the language type determination unit selectively performs the processing on a region having a sentence attribute. 26. A language type instruction input unit for inputting a language type instruction by a user, wherein the character recognition unit character-recognizes image information according to the language type input by the language type instruction input unit. 19. The image processing device according to claim 18, wherein: 27. The image processing apparatus according to claim 18, further comprising storage means for adding document identification information to the character recognition result and storing the result. 28. The image processing apparatus according to claim 18, further comprising storage means for storing the result of the character recognition and the input image information with document identification information added thereto. 29. The image processing apparatus according to claim 18, further comprising storage means for storing the result of the character recognition and the information obtained by compressing the input image information, with document identification information added thereto. 30. The image processing apparatus according to claim 18, further comprising storage means for storing the result of the character recognition and the layout information of the extracted area with document identification information added thereto. 31. An apparatus according to claim 27, further comprising output means for outputting a result of character recognition stored with the document identification information added thereto in accordance with the specification of the document identification information.
An image processing apparatus according to claim 1. 32. An apparatus according to claim 28, further comprising an output unit that outputs a result of character recognition and image information stored with the document identification information added thereto in accordance with the specification of the document identification information. Image processing device. 33. An image processing apparatus, comprising: output means for outputting, in accordance with the identification of the document identification information, a result of character recognition stored with the document identification information added thereto and image information compressed information. Item 30. The image processing device according to item 29. 34. An apparatus according to claim 30, further comprising output means for outputting a result of character recognition and layout information of an area stored with the document identification information added thereto in accordance with the specification of the document identification information. An image processing apparatus according to claim 1. 35. A control program for extracting a region for each attribute from an input image; a control program for determining a language type for each region; A computer-readable storage medium that stores a control program for a step of recognizing characters in response.