JPH01106263A - Document storage retrieving device - Google Patents

Abstract

PURPOSE:To eliminate a key word input work and to easily execute retrieval by providing the title device with a character recognizing device for segmenting characters in a document image, recognizing the segmented characters and converting the recognized characters into a character code string and a key word group extracting device. CONSTITUTION:The character recognizing device 7 segments character string areas 71-75 from a document image 70 read out by an image reader 2, successively recognizes the character images existing in the areas 71-75 while leaving the unclearness of division and outputs the codes of respective candidate characters. The device 7 including a reference character dictionary for KANA (Japanese syllabary) or KANJI (Chinese character) recognizes respective characters in each character and applies the recognized characters to a register 8 together with candidate characters. A key word group extracting device 10 checks the coincidence/discrepancy of all partial character code strings with length (n) included in a character code string 72 outputted from the register 8 and extracts key words.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an apparatus for converting documents into image data and storing and retrieving the data.

[Conventional technology]

As the amount of data and files increases, a large amount of documents are converted into electronic files and stored as electronic data on a high-density storage medium 2, such as an optical disk.

Conventionally, this type of device has been constructed as shown in FIG. When storing a document, the input document 1 is photoelectrically converted into a document image by the image reader 2, a keyword group according to the operator's arbitrary classification input from the keyboard 5 is attached, and the data is stored in a database 3 such as an optical disk serving as a storage medium. Store. On the other hand, when searching for a document, search conditions entered from the keyboard 5 (conditions according to the classification used when storing)
A document image attached with a keyword group that satisfies the following is searched, and the search results are displayed on the display device 6.

[Problem that the invention seeks to solve]

In such a device, for document 1 to be stored in advance,
It was necessary to attach an appropriate group of keywords based on the search situation. If you don't have a set of keywords, you won't be able to search the documents you have stored because there is no way to know where and what they have stored. Furthermore, when storing a document, the operator is required to perform a cumbersome task of creating and inputting a group of keywords. Even if it is sufficient for a document management device for personal use to memorize a group of input keywords, the problem is that it is difficult for another person to find them when searching.

The present invention has been made to eliminate such drawbacks, and eliminates the need for an operator to create a keyword group, automating the creation of a keyword group, and making it possible to easily perform a search even if the storer and searcher are different people. That is, it is an object of the present invention to provide a document storage and retrieval device that allows accurate storage and retrieval even when operators are changed.

[Means for Solving the Problems] The present invention includes an image reading device 2 that reads a document 1 and photoelectrically converts it into a document image, and a database 3 that stores the document image with data of a search keyword group. , a document storage and retrieval device comprising a search device 4 that searches and reads document images from the database 3 by inputting a group of search keywords, cuts out characters in the document image, recognizes them, and converts them into character code strings. The character recognition device 7 and the word dictionary 9 in which words and phrases for classification items are registered in advance, match the character code string with the words and phrases to extract keyword data regarding the document 1, and extract the document image. [Operation] When the operator inputs the document 1 using the image reading device 2, the document image is automatically assigned a keyword group by the keyword group extraction device 10 and stored in the database 3. When performing a search, even if the searcher is a different person, the desired document can be retrieved simply by inputting a group of general keywords for the field to which the document belongs.

〔Example〕

The present invention will be explained below based on the drawings. In FIG. 1, 1 is a document to be stored, and 2 is a document image of an electronic signal obtained by optically inputting a character image on document 1.

- an image reader as an image reading device that generates images; 3 is a database such as a magnetic disk or optical disk that stores electronic signals; 4 is a search device that searches for stored data in the database 3; 5 is a keyboard for inputting characters, etc.; 6 is a display device consisting of a CRT or the like.

7 is a character recognition device that cuts out each character data in the document image read by the image reader 2, recognizes it, and converts it into a character code; 8 is a character recognition device 7;
9 is a register that stores the character code string that is the output of .9 is a word dictionary as a ROM used for extracting keyword groups;
IO is a keyword group extraction device that extracts a keyword group from a character code string in a register 8 using a word dictionary 9.

The word dictionary 9 is a read-only memory in which systematic classification phrases, characters, etc. of terms related to the field of cooking, for example, are systematically registered in advance by a cooking expert. The character recognition device 7 has a built-in reference character dictionary and recognizes characters one by one.

Hereinafter, as an example, the operation will be described in the case where a document as shown in FIG. 2 is input into the apparatus.

First, the character recognition device 7 cuts out character string regions 71 to 75 from the image of the document read by the image reader 2, as shown in FIG. It sequentially recognizes each candidate character while preserving its character, and outputs the code of each candidate character. FIG. 4 shows an example of one character string area 72 obtained in this manner. In the figure, the first line is the first candidate character, and the second line is the second candidate character. The character recognition device 7 has a built-in dictionary of standard characters such as hiragana and kanji, and recognizes each character one by one and supplies them to the register 8 along with candidate characters.

Therefore, next, the keyword group extraction device 10 uses the register 8
For all partial character code strings of length n included in the character code string 72 output from , extract this as a keyword. This operation is performed for the length n in the range from the minimum length to the maximum length for character code strings 72 to 75, and character code strings 72 to 7
5, a group of keywords that match the registered classification terms are extracted. The keyword group extracted in this way is determined by the keyword group extracting device 10 as the classification item of document 1, and is used as the input document image.

is attached and stored in the database 3.

On the other hand, a search for the stored document 1 is performed by searching the database 3 for a document image attached with a keyword group that satisfies the search conditions for the desired field inputted from the keyboard 5.

That is, the operator inputs general classification terms and phrases related to the cooking field as keywords from the keyboard 5.

Then, the search device 4 detects a classification item that matches this keyword group from the database 3, reads document data 1 belonging to the matching classification item, and displays it on the display device 6.

The operator visually checks this displayed document, and if it is the desired document, print it out or take it in, and if it is different, input another group of keywords and search until the desired document is obtained.

〔Effect of the invention〕

As described above, according to the invention, in a document storage and retrieval device that is equipped with a search device that searches and reads document images from a database in which document images are stored by inputting a group of search keywords, a document storage and retrieval device is provided. A character recognition device that extracts and recognizes characters in an image and converts them into character code strings, a word dictionary in which words and phrases for classification items are registered in advance, and a character recognition device that extracts and recognizes characters in an image and converts them into character code strings, and a word dictionary that stores words and phrases for classification items in advance. A keyword group extraction device is provided which extracts keyword data and stores the keyword data attached to a document image in a database, and the keyword group attached to the document image to be stored is automatically extracted from the document image, Moreover, the automatically extracted keyword group always includes all words that can be keywords in the character strings included in the input document image, so the user does not have to create and input the keyword group when storing the document. Moreover, when searching for documents, no matter who searches for documents, documents that include a group of keywords that satisfy the specified search conditions are always searched based on the field-specific nature of the group of keywords attached to them.

[Brief explanation of the drawing]

FIG. 1 is an overall configuration diagram of the document storage and retrieval device of the present invention, FIG. 2 is a diagram showing an example of an input document, FIG. 3 is a diagram illustrating how a character area is cut out, and FIG. 4 is a character code FIG. 5, which is a diagram illustrating columns, is a configuration diagram of a conventional device. In the figure, 1 is the input document, 2 is the image reader, and 3
is a database, 4 is a search device, 5 is a keyboard, 6 is a display device, 7 is a character recognition device, 8 is a register, 9 is a word dictionary, 10 is a keyword group extraction device, 70 is a document image, and 71 to 75 are character areas It is. Agent Masu Oiwa A (and 2 others) Figure 1 Figure 2 Figure 3 Procedure amendment (voluntary) 1. Indication of the case Patent Application No. 1988-264888 3.
Column for detailed explanation of the invention to be amended by Representative Moriya Shiki 4 and Agent 5 of the person making the amendment. 6. Contents of the amendment (1) On page 2, line 6 of the specification, the phrase "to electronically file" should be amended to read "to electronically file." (2) In the 11th line of page 3 of the same book, the phrase ``It is an interlayer.'' has been amended to ``There is a problem.''that's all

Claims (1)

[Scope of Claims] An image reading device that reads a document and photoelectrically converts it into a document image; a database in which the document image is stored with data of a group of search keywords; A document storage and retrieval device comprising a retrieval device that searches and reads out the document image from the database, a character recognition device that cuts out and recognizes characters in the document image and converts them into a character code string, and a word for classification items. Extracting keyword data related to the document by comparing the character code string and the words and phrases with a word dictionary in which words and phrases are registered in advance, and extracting a keyword group to be added to the document image and stored in the database. A document storage and retrieval device comprising: