JP3026397B2 - Document search device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document search apparatus for performing a full-text search from a document coded by an optical character recognition apparatus.

[0002]

2. Description of the Related Art In recent years, with the practical use of optical character recognition devices, there has been an increasing demand for converting paper document information into character codes for retrieval. However, although the recognition rate of the optical character recognition device has improved considerably, the recognition rate is still poor for specific characters. Document that had to be proofread. In this proofreading work, although it is possible to specify the misrecognition position to some extent by Japanese language processing technology, it still needs to be manually performed, and one type in which the optical character recognition device is not used as the document input means of the full-text search device. That was the reason.

Hereinafter, a conventional document search apparatus will be described. FIG. 8 shows the entire configuration of a conventional document search apparatus. 8, reference numeral 801 denotes an optical character recognition device;
2 is a character code storage unit for storing a character code recognized by the optical character recognition device, 803 is a search condition input unit, 8
04 is a search means for searching according to the input search condition, 8
Reference numeral 05 denotes search result storage means for storing search results, and reference numeral 806 denotes search result output means for displaying and printing the search results.

The operation of the above-configured document search apparatus will be described below. First, the character code recognized by the optical character recognition device 801 is once stored in the character code storage unit 802. Next, according to the search condition input by the search condition input unit 803, the character code stored in the character code storage unit 802 is compared with the search unit 80.
The search is performed in step 4. Search results are stored in search result storage unit 805
, And the stored search results are displayed and printed by the search result output means 806.

As described above, even in the above-mentioned conventional document search apparatus, full-text search can be performed from a character-coded document using an optical character recognition apparatus.

[0006]

However, in the above-mentioned conventional document retrieval apparatus, there is a problem that retrieval is missed when performing retrieval because of an error in recognition of the optical character recognition apparatus.

An object of the present invention is to solve the above-mentioned conventional problems, and an object of the present invention is to provide a document retrieval apparatus capable of preventing omission of retrieval even when an optical character recognition apparatus makes a recognition error. It is.

[0008]

In order to achieve the above object, a document retrieval apparatus according to the present invention comprises: an optical character recognition apparatus for performing character encoding from printed characters; and a document encoded by the optical character recognition apparatus. Character code storage means for storing
Search condition input means for inputting search conditions, division error rule storage means for storing, as a rule, an example where a division error is likely to occur when the optical character recognition device divides a character, and a table of characters
An example where rules are likely to cause recognition errors due to similar
Typographical error storage means for storing and inputting the search condition
The division error rule case for the search condition input by the means
The condition is expanded using the payment method, and the spelling error rule
A search condition expansion means for performing condition extended with storage means, a search means for searching the character code storage means by expanded search condition in the search condition expansion means, and the search result storing means for storing the result of the search means, the search And a search result output means for displaying and printing the result.

The present invention further comprises a notation error rule storing means for storing, as a rule, an example in which recognition errors are likely to occur due to similar character notation of the optical character recognition device. After performing the condition expansion, the search condition expansion means expands the condition using the notation error rule storage means.

[0010]

According to the present invention, when the character code of the search condition is a character code which is likely to be erroneously recognized by the optical character recognition device, the search omission can be reduced by extending the search condition from the recognition rule. it can. The optical character recognition device performs division processing to determine the area of one character. However, a character that is far from the bias is determined to be another character, and a division error occurs. In such a case, since it is possible to determine in advance from the form of the character, this is registered, and if this character is included in the search condition, the character that caused the division error is also added. The extension can prevent search omission due to a division error.

According to the present invention, even when notation cannot be recognized by an optical character recognizing device due to similar notation, it is possible to determine in advance the similar notation from the character form. Is registered, and if this character is included in the search condition, search omission can be prevented by expanding the condition.

[0012]

【Example】

(Embodiment 1) Hereinafter, a first embodiment of the present invention will be described with reference to FIG. In FIG. 1, reference numeral 101 denotes an optical character recognition device for converting a print character to a character code;
Reference numeral 2 denotes a character code storage unit for storing a character code recognized by the optical character recognition device 101; 103, a search condition input unit for inputting search conditions; and 104, an optical character recognition device 1.
A division error rule storage unit 105 stores, as a rule, an example in which 01 is likely to cause a character division error. The division error rule storage unit 105 uses the division error rule storage unit 104 for the search condition input by the search condition input unit 103. Search condition expansion means for expanding the search condition, 106 is a character code storage means 1
02, a search means for performing a search based on the search condition created by the search condition expansion means 105, and 107 a search means 10
Search result storage means 108 for storing the search result obtained in step 6;
Is a search result output unit for displaying the search results of the search result storage unit 107 on a screen or printing the search results.

The operation of the above-configured document retrieval apparatus will be described below. First, a printed document is recognized as a character code by the optical character recognition device 101,
It is stored in the character code storage means 102. The search condition is input from the search condition input unit 103, and the search condition expansion unit 105 decomposes the word of the search condition into character units. Next, an inquiry is made to the division error rule storage means 104 as to whether each of the decomposed characters corresponds to one of the division error rules stored in the division error rule storage means 104. In the case of the division error rule, the search condition expansion unit 105 expands the condition by adding it to the search condition according to the rule. Further, the search condition expansion unit 10 searches the character code stored in the character code storage unit 102 by the search unit 106.
The search is performed according to the search condition extended in step 5. The search result is stored in the search result storage unit 107, and is output to the search result output unit 108 when displaying and printing.

FIG. 2 shows a case where the optical character recognition apparatus uses a character 201.
Is shown as an extended example 202 in the case of recognizing the divided error rule into two parts, that is, the partial error rule and the partial error rule.
Indicates a rule to be input to. For example, the character "ka" is apt to be erroneously recognized as two characters "a" and "hi" because the characters are separated from each other.

FIG. 3 is a flowchart showing an extension procedure of the search condition extension means 105 of FIG. When the search condition is input, the process starts (step 301), and first, the word of the search condition is decomposed into characters (step 30).
2). Next, it is determined whether or not each decomposed character is included in the division error rule (step 303). If it is included in the rule, the character is extended according to the division error rule and added to the search condition as an OR condition. (Step 3
04). If it is not included in the rules, do nothing and end. (Step 305)

As described above, according to the present embodiment, even if a character is likely to cause a division error in the optical character recognition device, the retrieval condition can be expanded by the division error rule, thereby preventing the retrieval omission. .

(Embodiment 2) Next, a second embodiment of the present invention will be described with reference to FIG. In FIG.
401 is an optical character recognition device, 402 is a character code storage unit, 403 is a search condition input unit, 404 is a division error rule storage unit, 405 is a search condition expansion unit, 406 is a notation error rule storage unit, 407 is a search unit, 408 is a search result storage means, and 409 is a search result output means. The difference from the configuration of FIG. 1 is that a notation error rule storage unit 407 for storing, as a rule, an example in which recognition errors are likely to occur due to similar character notations is added.

The operation of the above-configured document retrieval apparatus will be described below. First, a printed document is recognized as a character code by the optical character recognition device 401.
The character code is stored in the character code storage unit 402. The search condition is input from the search condition input unit 403, and the search condition expansion unit 405 decomposes the word of the search condition into character units.
Next, the division error rule storage unit 404 is inquired about whether or not each of the decomposed characters corresponds to the division error rule. If the character is included in the division error rule, the search condition expansion unit 405 adds it to the search condition according to the rule. Extend search conditions.
The operation up to this point is the same as in the first embodiment. The difference is that thereafter, it is determined whether or not there is a different character code in the same notation for each character extended by the division error rule. For example, the katakana "b" is almost the same as the kanji "kuchi", so it is easy to misrecognize it, and "a" and "a" are easy to misrecognize because only the character size is different. Such a character code that is apt to be erroneously recognized is registered as a notation error rule in the notation error rule storage unit 406, and the condition is extended. Further, the search unit 407 searches the character code stored in the character code storage unit 402 according to the search condition expanded by the search condition expansion unit 405. The search result is stored in the search result storage unit 408, and is output to the search result output unit 409 when displaying and printing.

FIG. 5 shows an example of a typographical error rule stored in the typographical error rule storage means 406. An extended example 501 when the notation is similar, and an extended example 50 when the character size is different.
2 shows an extended example 503 in the case of an error in a cloud point. For example, the katakana “ka” is almost the same as the kanji “power (Chikara)” and is easy to misrecognize. Because it is different, it is easy to mistakenly recognize.

FIG. 6 is a flowchart showing an extension procedure of the search condition extension means 405 of FIG. When the search condition is input, the process starts (step 601). First, the words of the search condition are decomposed into characters (step 60).
2) Next, it is determined whether or not each decomposed character is included in the division error rule (step 603). If it is included in the rule, it is extended according to the division error rule and added to the search condition as an OR condition (step 6).
04). The operation up to this point is the same as in the first embodiment. The difference is thereafter, and it is determined whether each character including the character extended by the division error rule is included in the writing error rule (step 605). If it is included, it is extended according to the notation error rule and added as an OR condition (step 606). If it is not included in the rule, the process ends without doing anything (step 60).
7).

FIG. 7 is a diagram showing how the condition expansion is performed for the search condition of the character "processing". First, "processing" is set as a search condition (step 701), and then a division error rule is applied.
Since “addition” is easily misrecognized as “power” and “mouth”, the search condition is extended to “processing” and “power mouth” (step 70).
2). Furthermore, when the spelling rule is applied, "ko" has the same notation as "e", "power" has the same notation as "ka", and "mouth" has the same notation as "b". Therefore, this rule is applied to each of “machining” and “strengthening” (step 703). As a result, the search condition of "processing" is extended to ten characters.

As described above, according to this embodiment, even when there is a recognition error for a similar notation when a character is divided by the optical character recognition device, the search condition is extended by the notation error rule. This can prevent search omission.

[0023]

As described above, the present invention provides a division error rule storage means, a notation error rule storage means, and a search condition expansion means for expanding a search condition to compensate for a recognition error of an optical character recognition device. Is added, it is possible to prevent search omission due to recognition errors, and to realize an excellent document search device.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a document search device according to a first embodiment of the present invention.

FIG. 2 is a list showing an example of a division error rule in the first embodiment.

FIG. 3 is a flowchart illustrating a procedure for expanding a search condition in the first embodiment;

FIG. 4 is a block diagram illustrating a configuration of a document search device according to a second embodiment of the present invention.

FIG. 5 is a list showing an example of a spelling error rule in the second embodiment.

FIG. 6 is a flowchart showing a procedure for expanding search conditions in the second embodiment.

FIG. 7 is an arrow chart showing a specific example of search condition expansion in the second embodiment.

FIG. 8 is a block diagram showing a configuration of a conventional document search device.

[Explanation of symbols]

 Reference Signs List 101 optical character recognition device 102 character code storage means 103 search condition input means 104 division error rule storage means 105 search condition expansion means 106 search means 107 search result storage means 108 search result output means 401 optical character recognition device 402 character code storage Means 403 Search condition input means 404 Division error rule storage means 405 Search condition expansion means 406 Notation error rule storage means 407 Search means 408 Search result storage means 409 Search result output means

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor ▲ Tsuru ▼ Ken Hayashi 1006 Kazuma Kadoma, Osaka Prefecture Inside Matsushita Electric Industrial Co., Ltd. (56) References JP-A-4-92971 (JP, A) JP-A-61-221889 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06F 17/30

Claims (1)

(57) [Claims] 1. An optical character recognition device for performing character encoding from printed characters, character code storage means for storing a document encoded by the optical character recognition device, and search condition input means for inputting search conditions. When the optical character recognition device divides a character, the division error rule storing means for storing an example in which a division error is likely to occur as a rule is similar to the character notation.
Examples where recognition errors easily occur because of
Writing error rule storage means, and the search condition input means
The division error rule storage means for the input search condition
To expand the condition using
Search condition expansion means for performing a condition expansion using a stage , search means for searching from the character code storage means according to search conditions expanded by the search condition expansion means, and search result storage means for storing results of the search means , Show search results
A document search device comprising a search result output unit for printing.