JPH01169689A - Character recognition device - Google Patents

Abstract

PURPOSE:To facilitate to input a fixed form phrase comprising of plural character strings by providing a correspondence means to correspond a dictionary character pattern and the character string of one character or more and an output means to output a character string corresponding to a selected dictionary character pattern. CONSTITUTION:A conventional character recognition device is equipped with a fixed form phrase dictionary 10 of one pattern plural character strings correspondence and an output means to output a character string corresponding by a recognition result. When a pattern indicating a fixed form phrase is hand- written, the character recognition device selects a dictionary pattern which most looks like the hand-written pattern out of general/fixed form phrase dictionaries 8 and 10. Since the character string is corresponded to the fixed form phrase dictionary 10, it is outputted by the output means. Thus, the fixed form phrase comprising of plural characters can be inputted only by hand-writing one pattern.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention belongs to the technical field of handwritten character recognition, and particularly relates to fixed phrase input suitable for handwritten character input devices.

[Conventional technology]

In the conventional handwriting input device, as described in Japanese Unexamined Patent Publication No. 58-123177, characters are recognized based on the correspondence of 1 handwritten pattern→] character.

[Problem that the invention seeks to solve]

The above-mentioned conventional technology takes into consideration the ease of inputting standard directions, but even for standard directions that are frequently used in sentences, there is a problem in that it is necessary to input handwriting patterns for the number of characters included in the standard directions. Ta.

An object of the present invention is to make it easier to input a fixed format consisting of a plurality of character strings.

[Means for solving problems]

The above object can be achieved by providing a conventional character recognition device with a fixed phrase dictionary that can handle one pattern → multiple character strings, and an output means for outputting a character string corresponding to the recognition result.

[Effect]

When you handwrite a pattern that indicates a certain standard direction, the character recognition device selects a dictionary pattern that is most similar to the handwritten pattern as a general/general pattern.
Select from the fixed phrase dictionary. Since character strings are associated with the fixed phrase dictionary, these are outputted by the output means. Now, you can input a standard format consisting of multiple characters just by handwriting a single pattern.

〔Example〕

Hereinafter, as an example of the present invention, recognition of a fixed format orientation and,
A handwritten character string recognition device in which a user can register a fixed format will be explained using FIGS. 1 to 4. FIG.

FIG. 1 is a diagram showing its overall configuration.

First, we will explain the components necessary for recognizing characters and fixed formats. The tablet 1 reads input strokes as sample point coordinates for each unit time and outputs them to the signal line 101. The stroke data compression device 2 has N stroke data for each rotor.
A point is approximated, and an approximate pattern for one character is output to the signal line 102. The matching device 3 selects the dictionary pattern most similar to the six-point approximation pattern received from the signal line 102 from the dictionary 8 via the signal line 103, and outputs its recognition code to the signal line 104. In the dictionary 8, a six-point approximation pattern of one handwritten character is registered in association with one recognition code. Here, if this recognition code is only a character code, the configuration described above is the same as the conventional method. However, in this embodiment, the recognition code includes a fixed format code in addition to the character code. The standard pattern matching device 4 determines whether the recognition code received from the signal line 104 is a character code or a standard phrase code, and if it is a character code 1-, it outputs it as is to the signal line 110, and if If it is a phrase code, a string of characters corresponding to the phrase code is read out from the phrase dictionary 10 via the signal line 105 and outputted to the signal line 110. In the fixed phrase dictionary 10, one fixed phrase character code is registered in association with a character code string of characters included in the fixed phrase.

The second part is the configuration necessary for character/fixed phrase recognition.

In this embodiment, a code input device 5 such as a keyboard is also connected to the signal line so that character codes can be directly input without recognizing handwritten patterns.

Next, the components necessary for the user to register characters/fixed phrases will be explained.

The fixed phrase check device 6 receives 'ARu' from the signal line 106.
It receives the character code string to be registered, and determines whether to register one character or a fixed phrase based on this length. If it is a single character registration, a character code is output to the signal line 107, and if it is a fixed phrase registration, it outputs the character code to the signal line 107. For example, a code that can be distinguished from character codes is automatically generated as a fixed phrase code, and this code is output to the signal line 107, and further the character code string received on the signal line 106 is output.

The dictionary registration device 7 receives the character/standard orientation code from the signal line 107, and also receives the tablet 1-. The approximation patterns handwritten on the lines 1 and 2 and approximated by the stroke data compression device 2 are received and registered in the dictionary 8 via the signal line 108 so as to be associated with each other. The fixed phrase registration device 9 receives a character code string of characters forming a fixed phrase from a signal line 106, and receives a fixed phrase code from a signal line 107.
These are registered in the fixed phrase dictionary 10 via the signal line 109 so as to be associated with each other.

FIG. 2 shows an example of registration of the dictionary 8 and the fixed phrase dictionary 10. For single characters, as in the past, the dictionary 8 associates approximate patterns of handwritten patterns with their character codes. For fixed phrases, the dictionary 8 associates approximate patterns of handwritten patterns with corresponding fixed phrase codes.

Further, in the fixed phrase dictionary 10, fixed phrase codes are associated with character code strings forming fixed phrases.

FIG. 3 shows a procedure for recognizing characters or fixed phrases. First, in the same process 1 as before, the tablet 1 in FIG.
, the stroke data pressure 1 device it2, the matching device 3, and the dictionary 8 recognize the input pattern and obtain a recognition code. Next, in process 2 in FIG. 3, it is determined whether the recognition code obtained by the fixed phrase matching device 4 is a character code or a fixed phrase code. The character code is usually
Although expressed as IWORD, the code range used for the character code 1- is limited. Therefore, if a fixed phrase code is assigned to a code range different from the character code, the character/fixed phrase code can be easily determined. Further, the fixed phrase dictionary 10 is searched and the presence/absence of the phrase is determined. Alternatively, it is also possible to determine whether to specify character recognition mode or fixed phrase recognition mode from the outside.

If it is determined to be a fixed phrase code in processing 2, in processing 3, the fixed phrase dictionary 10 is checked with the obtained fixed phrase code,
Read and output the fixed phrase character code string. Hello again,
If it is determined in process 2 that it is not a fixed phrase code, the recognition code is a character code, so in process 4 a recognition code is output.

FIG. 4 shows the procedure for registering characters or fixed phrases by the user.

First, in process 1, the character code string of the character or fixed phrase to be registered is input. This may be input directly from the code input device 5 shown in FIG. 1, such as a keyboard, or may be input by hand using the handwriting pattern recognition described above.

Next, in process 2 of FIG. 4, the length of the input character code string is checked, and if it is one character long, it is determined that one character is registered, and if it is two or more characters, it is determined that it is a fixed phrase registration.

This determination can also be made by externally specifying whether the mode is one-character registration mode or fixed phrase registration mode.

If it is determined in process 2 that the phrase has been registered, a phrase code is automatically generated and used as a recognition code in process 3.

By limiting the range of this code and keeping it separate from the range of character codes, the character /
It is easy to determine the fixed format code.

Next, in process 4 in FIG. 4, the fixed phrase registration device 9 in FIG. 1 registers the fixed phrase dictionary 10 so that the character string obtained in process 1 and the fixed phrase code generated in process 3 correspond. Hello again,
If it is determined in process 2 that one character has been registered, in process 5 the character code obtained in process 1 is used as the recognition code.

Next, in process 6, a handwritten pattern of a character to be registered is inputted from the tablet 1 shown in FIG. 1, and the stroke data compression device 2 performs N-point approximation to obtain an approximate pattern. Finally, in step 7 of FIG. 4, the recognition code obtained in step 3 or step 5 is registered in the dictionary 8 so that the approximate pattern obtained in step 6 corresponds.

〔Effect of the invention〕

According to the present invention, when creating a document using a handwritten character input device,
The hassle of inputting frequently used fixed phrases will be eliminated and efficiency will be improved. Quantitatively, if a fixed phrase and its pattern are registered once, in the case of an N-character fixed phrase, the effort is reduced to 1/H.

Furthermore, during recognition and registration by the user, the user does not have to specify whether it is a single character or a fixed phrase, improving usability.

[Brief explanation of the drawing]

FIG. 1 is an overall configuration diagram of an embodiment of the present invention, FIG. 2 is a diagram showing registered contents of a dictionary 8 and a fixed phrase dictionary 10, and FIG. 3 is a character/fixed phrase recognition flowchart 1-5 4. Illustration is text/
It is a flowchart for registering fixed phrases. 1...Tablet, 2...Stroke smoothing device,
3. Matching device, 4. Fixed phrase matching device,
5... Court input device, 6... Fixed phrase check device,
7 Dictionary registration device, 8... Dictionary, 9... Fixed phrase registration device 2nd day, 5 mouth

Claims (1)

[Scope of Claims] 1. A character recognition device comprising a plurality of dictionary character patterns, a character pattern input device for inputting character patterns, and a matching device for selecting the dictionary character pattern most similar to the character pattern, A character recognition device comprising: a correspondence means for associating the dictionary character pattern with a character string of one or more characters; and an output means for outputting a character string corresponding to the selected dictionary character pattern. 2. In claim 1, the corresponding means:
A character recognition device characterized in that it is a dictionary in which one dictionary character pattern is registered in correspondence with one or more character strings. 3. In claim 1, the corresponding means:
It is characterized by being composed of a character dictionary in which a plurality of dictionary character patterns are registered in association with one code, and a fixed phrase dictionary in which a plurality of character strings of one or more characters are registered in association with the code. character recognition device. 4. In claim 3, the output means:
A character recognition system comprising: a code output means for outputting a code corresponding to a dictionary character selected by the character dictionary; and a character string output means for outputting a character string corresponding to the code from the fixed phrase dictionary. Device. 5. In claim 3, a range of codes corresponding to a dictionary character pattern in the character dictionary that is associated with a character string of one character, and a range of codes that are associated with a character string of two or more characters. A character recognition device characterized in that code ranges corresponding to dictionary character patterns do not overlap. 6. In claim 3, if the one dictionary character pattern is associated with a character string of one character, in the character dictionary, the code associated with the dictionary character is the character string. A character recognition device characterized in that the character code is 7. The character recognition device according to claim 3, wherein a dictionary character pattern associated with a character string of one character is not registered in the fixed phrase dictionary. 8. Character recognition according to claim 4, characterized in that in the case of a dictionary character pattern associated with a character string of one character, the character of the code output by the code output means is output. Device. 9. In claim 4, in the character dictionary,
Outputting a character string by the character string output means only when a code associated with a dictionary character pattern is within the range of a code corresponding to a dictionary character pattern associated with a character string of two or more characters. A character string recognition device featuring: 10. In claim 4, a mode specifying means is provided for specifying a one character input mode and a character string input mode of two or more characters, and when the code is in the one character input mode, the code is output by the code output means. A character recognition device characterized by outputting characters of a code. 11. In claim 4, a mode specifying means is provided for specifying a one character input mode and a character string input mode of two or more characters, and only when the character string input mode is two or more characters, the character string is output. A character recognition device characterized by outputting a character string by a means.