JPH06325201A - Character recognizing device - Google Patents

Abstract

PURPOSE:To make it possible to recognize characters with high precision corresponding to a specified document transmitted from each department by switching a knowledge dictionary according to transmission originating sides also when different kinds of documents are entered in a slip of the same format. CONSTITUTION:In a reception part 11, a received facsimile signal is converted into the image signal of a document, further, transmission originating side data are extracted from the facsimile signal, the image signal of the document is outputted to a character recognition part 12 and transmission originating data are outputted to a knowledge dictionary selection part 13. In the character recognition part 12, characters are extracted one by one from document image data, a recognition processing is performed by using a recognition dictionary (not shown in figure) and the character code of the result is outputted to a knowledge processing part 16. In a knowledge dictionary selection part 13, a knowledge dictionary corresponding table is referred to in accordance with transmission originating data, first or second knowledge dictionary 14, 15 is selected and a state retrievable from the knowledge processing part 16 is made. In the knowledge processing part, a word is extracted from the character code name string of a recognition result and a collation is performed by using the selected knowledge dictionary 14 or 15.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition device connected to a communication line and capable of highly accurate recognition.

[0002]

2. Description of the Related Art Conventionally, as a character recognition device connected to a communication line and capable of operating efficiently, for example, Japanese Patent Application No. 03-18
There is a character recognition device included in the information processing device 1072. This information processing apparatus extracts an image signal of a document and data representing a sender from a signal transmitted through a communication line, detects a character area from the received image signal, and detects a character in the detected character area. Character recognition is performed, the recognized character recognition result is translated, and the translation result is returned to the received sender.

As a character recognizing device capable of recognizing with high accuracy, there is a character recognizing device shown in Japanese Patent Application No. 63-029098, which recognizes characters one by one and
A word is extracted from a recognition result string for each character, the extracted word is collated with a word dictionary prepared in advance, and the result of the word collation is output as a recognition result.

[0004]

However, the above-mentioned conventional character recognition device connected to a communication line and operated is generally intended to efficiently operate a small number of devices in a plurality of departments in a company or the like. In many cases, it was therefore the case to receive and process different documents from different departments. Therefore, the conventional character recognition device has to have a general-purpose structure, and therefore cannot generally support high-precision recognition because it cannot correspond to a specific document. Therefore, a document can be efficiently input. There was a problem that I could not do it.

For example, as shown in FIGS.
When the form document images, which are written on the form of the same format but have different types and contents of documents due to different senders, are recognized, the conventional character recognition device uses the same format. Since it is written on the form, the same word dictionary must be used for the word matching in the character recognition processing, and a satisfactory word matching cannot be performed, resulting in a problem that high-accuracy recognition cannot be performed.

An object of the present invention is to provide a character recognition device capable of highly accurate recognition corresponding to a specific document transmitted from each department.

[0007]

In order to solve the above-mentioned problems, the present invention provides a character recognition device for recognizing a document transmitted from a communication line, and a receiving unit for receiving data representing an origin together with an image signal. A character recognition unit that recognizes each character of the received document image, a knowledge dictionary selection unit that selects a knowledge dictionary according to the received source data, and a recognition result of each character. A knowledge processing unit that performs knowledge processing on the column using the selected knowledge dictionary.

[0008]

According to the character recognition apparatus of the present invention, since it is configured as described above, the knowledge dictionary can be switched according to the sender even when different types of documents are written on the same form. Therefore, knowledge processing can be performed using the knowledge dictionary corresponding to the transmitted document, and highly accurate recognition can be performed.

[0009]

1 is a block diagram showing an embodiment of a character recognition device of the present invention. The character recognition device 10 includes a reception unit 11, a character recognition unit 12, a knowledge dictionary selection unit 13, and a first knowledge dictionary 1.
4, a second knowledge dictionary 15, a knowledge processing unit 16, and an output unit 17. FIG. 2 is a diagram showing an example of a document image,
Reference numeral 20 is a book image document, and 21 is a character entry frame. FIG. 3 is a diagram for explaining the knowledge dictionary. FIG. 4 is a diagram for explaining the knowledge dictionary correspondence table, where 40 is a knowledge dictionary correspondence table, 41 is a source telephone number, and 42 is a knowledge dictionary number.

An embodiment of the present invention will be described in detail below with reference to FIG. First, the facsimile signal S transmitted through the communication line is input to the receiving unit 11 in the character recognition device 10. In the receiving unit 11, the received facsimile signal is converted into a document image signal, the source data is extracted from the facsimile signal, the document image signal is output to the character recognition unit 12, and the extracted source data is extracted. Is output to the knowledge dictionary selection unit 13.

The character recognition unit 12 extracts a character by character from the document image data input from the reception unit 11, and performs a recognition process on the character by character using a recognition dictionary (not shown). Then, the character code name of the recognition result is output to the knowledge processing unit 16.

In the knowledge dictionary selecting section 13, the receiving section 1
The knowledge dictionary correspondence table is referred to according to the source data input from 1 to select either the first knowledge dictionary 14 or the second knowledge dictionary 15, and the knowledge processing unit 16 searches the selected dictionary. Make it possible.

In the knowledge processing unit 16, the character recognition unit 1
A word is extracted from the character code name string of the recognition result input from 2 and the word is matched using the knowledge dictionary selected by the knowledge processing selection unit 13 and the word character string of the matching result is output as the recognition result. It is output to the unit 17. The word extraction is performed by dividing the character code name sequence of the recognition result input from the character recognition unit 12 at the delimiters of “,”, “.”, “” And the places where the character type changes. To do.

The output unit 17 is a terminal for outputting the word character string of the recognition result input from the knowledge processing unit 16 to the outside, and is connected to an external storage device, communication line, other information processing device, or the like. Is.

Next, a specific example of this embodiment will be described with reference to FIGS. FIG. 2 is a diagram showing an example of the transmitted facsimile document image, which is shown in FIGS.
(B) are, for example, different types of documents created in different departments. 3 is a diagram for explaining the knowledge dictionary. FIG. 3 (A) is a first knowledge dictionary 15 in which addresses, names, contract-related words, etc. are registered, and FIG. 3 (B) is a general word is registered. It shows an example of registered words of the generated second recognition dictionary 16. FIG. 4 is a diagram showing an example of a knowledge dictionary correspondence table in the knowledge dictionary selection unit 13, where 41 is a source department telephone number and 42 is a knowledge dictionary number corresponding to a document transmitted from the department in advance. It has been done.

First, when a document as shown in FIG. 2A is input from one department connected by a communication line,
In the receiving section 11, a document image as shown in FIG. 2A is sent to the character recognizing section 12, and a source telephone number (1005 in this embodiment) which is the calling department data extracted from the received facsimile signal. Knowledge dictionary selection unit 13
Output to. In the knowledge dictionary selecting unit 13, the receiving unit 1
From the department telephone number 1005 input from No. 1, the knowledge dictionary number corresponding to the type of the document sent by the department is detected using the knowledge dictionary correspondence table shown in FIG. 4, and (1 in this embodiment), the dictionary number concerned. The knowledge dictionary 1 as shown in FIG. 3A corresponding to is selected so that the knowledge processing unit 16 can retrieve it.

When a facsimile signal as shown in FIG. 2 (B) is input from another department (the department telephone number is 1006), the receiving unit 11 displays the document image as shown in FIG. 2 (B). Is output to the character recognition unit 12, and the department telephone number 1006 of the department that is the source of the document is output to the knowledge dictionary selection unit 13. The knowledge dictionary selection unit 13 uses the knowledge dictionary correspondence table shown in FIG.
A knowledge dictionary number corresponding to the type of the document transmitted by the department indicated by 06 is detected, and the knowledge processing unit 16 performs a process for searching the corresponding knowledge dictionary 2 as shown in FIG. 3B.

Therefore, in the knowledge processing section 16, FIG.
When a document image with contract information as shown in (A) is recognized using the knowledge dictionary 1, and a document image with a general document as shown in FIG. 2B is input. Knowledge processing is performed using the knowledge dictionary 2.

Although the character recognition device of the present invention has been described in detail above, the present invention is not limited to the present embodiment and can be suitably implemented within the scope of the present invention.
For example, in the present embodiment, there are two types of knowledge dictionaries, but the number of knowledge dictionaries is not limited to this, and the knowledge dictionaries of three or more types may be provided. Further, the knowledge dictionary is a contract-related word dictionary and a general-word dictionary, but the knowledge dictionary is not limited to this, and a dictionary having an arbitrarily suitable structure corresponding to each document may be used.

[0020]

As described in detail above, the character recognition apparatus of the present invention is a character recognition apparatus connected by a communication line and used by a plurality of departments. Even when written and transmitted, the knowledge dictionary can be switched according to the sender,
It becomes possible to perform knowledge processing using a knowledge dictionary according to the document transmitted from each department. Therefore, in each department, the communication line is specialized as much as a character recognition device in its own department. It is possible to use the character recognition device connected by, so that highly accurate recognition is possible and an efficient document input system can be constructed.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a form image used in this embodiment.

FIG. 3 is a diagram illustrating a knowledge dictionary.

FIG. 4 is a diagram illustrating a knowledge dictionary correspondence table.

[Explanation of symbols]

 10 Character Recognition Device 11 Reception Unit 12 Character Recognition Unit 13 Knowledge Dictionary Selection Unit 14 Knowledge Dictionary 1 15 Knowledge Dictionary 2 16 Knowledge Processing Unit 17 Output Unit

Claims (1)

[Claims] 1. A character recognition device for recognizing a document transmitted from a communication line, and a receiving unit for receiving data representing a sender together with an image signal and a character for each character of the received document image. A character recognition unit, a knowledge dictionary selection unit that selects a knowledge dictionary according to the received source data, and a knowledge processing using the selected knowledge dictionary for the character recognition sequence of each character. A knowledge processing unit that
A character recognition device comprising: