JP3299875B2 - Character processor - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character processing device for inputting and editing characters by handwriting input or OCR (optical reader).

[0002]

2. Description of the Related Art Conventionally, it is assumed that handwriting input is made possible by collating information of a character stroke input by handwriting through a coordinate input device such as a tablet with a handwriting recognition dictionary prepared in advance, and adding one to the input information. In some cases, the closest one was output as the matching result. In addition, by presenting some items close to the input information as candidates and allowing the user to select them, the low recognition rate is covered.

[0003] However, it is very troublesome to input characters having a large number of strokes, and it is very inconvenient to re-enter the characters if the character cannot be recognized. Will decrease.

As a means for avoiding these problems, there is a method of inputting a radical in a handwriting input frame, displaying a list of kanji having the radical, and selecting and inputting a desired kanji from the kanji.

There is also known a method of simplifying an input using a handwritten user dictionary. This is to register a predetermined character in the handwriting input pattern. As shown in FIG. 16, when it is desired to write the character of "gi" in abbreviated form, by registering the abbreviated handwritten pattern in the handwritten user dictionary in association with the written pattern, the registered pattern is handwritten in the next time. By
The character of "gi" is automatically collated and output. Also, a plurality of character strings can be registered in correspondence with one registration pattern. This example is shown in FIG.

[0006]

However, when inputting a radical and displaying a list of kanji having the radical,
When "kiken" or "sanzui" is input, a very large number of candidates are displayed, and it is very difficult to find and specify a desired kanji from them.
In addition, there is no problem if the character is a kanji that allows the radical of the character to be immediately understood, but this function could not even be used for a character whose radical cannot be determined.

Further, when the above-described handwritten user dictionary is used, one handwritten pattern is always required for one specified character. In addition, even when the character strings are registered in the handwritten user dictionary in correspondence with each other, another dictionary must be newly registered for a character string that differs even by one character. Similarly, since the number of registered words increases, There have been problems that the dictionary capacity increases and that dictionary management becomes complicated.

In view of these problems, in the present invention, the dictionary capacity is reduced by registering a part of characters in a handwritten dictionary. By registering a part of the character, the part of the character is the same, and even if the part is the same (for example, “gi” or “gi”), it is possible to eliminate the waste of registering the same part. it can. Therefore, dictionary management can be facilitated.

[0009]

According to claim 1 of the present invention,
In a character processing device for comparing an input pattern input from a handwriting input unit such as a tablet with a character pattern registered in advance in a handwritten character recognition dictionary to obtain a recognition result for the input pattern, a character processing device that receives the input pattern from the input unit. Registering the input pattern as a partial pattern of a specific character in the handwritten character recognition dictionary, and storing a pattern registered by associating the input pattern with the partial pattern by the partial pattern registering means. A pattern registration dictionary; a partial pattern search unit for determining whether or not a partial pattern registered in the partial pattern dictionary is input by handwriting in an input pattern input from the input unit; All or a part of the input pattern is Based on the result, a partial pattern replacement means for replacing the data read from the partial pattern registered dictionary, for the changed pattern by said partial pattern replacement unit, further partial pattern search
Search using the partial pattern replacement means.
Repeat until all replacements are completed
After, against the handwriting recognition dictionary, to solve the above problem by outputting the recognition result. In addition, the claims of the present invention
In item 2, the partial pattern is registered in the partial pattern registration dictionary
Partial patterns are input patterns that have already been registered
By making it possible to register hierarchically using
Solution to the Problems

[0010]

[0011]

[0012]

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing an apparatus configuration of an embodiment of the present invention. 1 is a handwriting input unit. Usually, a coordinate input device such as a tablet is used. Reference numeral 2 denotes handwritten character recognition means. The handwritten character recognizing means 2 is a means for collating input data input from the handwriting input means 1 with a handwritten character recognition dictionary 3 described later. The handwritten character recognition dictionary 3 is a dictionary for comparison with data input from the handwriting input unit 1.
The data is stored in the form of stroke information, image information, or the like.

Reference numeral 4 denotes a partial pattern registration unit. The partial pattern registration unit 4 is a unit that registers the input data together with the corresponding partial pattern in the partial pattern registration dictionary 5. The partial pattern search means 6 is a handwriting input means 1
This is a means for searching whether or not the input data registered in the partial pattern registration dictionary 5 exists in the input data input from. The partial pattern replacement means 7 is a means for reading data from the partial pattern registration dictionary 5 and replacing it when the partial pattern search means 6 searches for a partial pattern in the input pattern. Reference numeral 8 denotes an output unit that outputs a result recognized by the handwritten character recognition unit 2.

The processing flow will be described below in order. First, the partial pattern registration processing will be described. FIG. 2 is a flowchart showing the flow of the partial pattern registration process. FIG. 3 is a diagram showing a screen during processing. Step S1
In step 1, the user inputs partial pattern data of a character to be registered in the input means 1. FIG. 3 shows an example in which the character “dragon” is input. The data input in step S12 is stored in the dictionary. Subsequently, an input pattern of a character to be registered is input to the input means 1 in step S13.
In step S14, it is registered in pairs with the data stored in step S12. In step S15, the order is changed in the order of the number of strokes of the input pattern data in the dictionary. If there is nothing to register further in step S16, the registration process ends, and if there is, the process returns to step S11 to continue the registration work. FIG. 4 shows the structure of the dictionary for which registration has been completed.

The processing when an input is made to the input means 1 will be described with reference to the flowchart of FIG. The coordinates of the character input in step S1 are converted into data in step S2. Then, in step S3, reference is made to the partial pattern registration dictionary which is the central part of the present invention, normal handwritten character recognition processing is performed in step S4, and data is output from the output means in step S5. Hereinafter, the step S3 will be described in detail.

The processing will be described based on the flowchart of FIG. 6 and FIG. In step S21, it is checked whether or not there is a registration in the registration dictionary (partial pattern registration dictionary 5). If not, the process ends. That is, only normal handwritten character recognition is performed.

If there is a registration, the designated input pattern data registered in step S22 is referred to, and in step S23, the pattern data input by handwriting by the collating means is compared with the data. In step S24, the partial pattern search means 6 determines whether the specified input pattern data is included in the pattern data input by handwriting. If not included, the process returns to step S21 to refer to the next pattern data. If it is included, the process proceeds to step S25. In step S26, the specified input pattern data in the input pattern data is replaced with the specified partial pattern data of the character by the partial pattern replacement means 7.

Referring to FIG. 7, how an actual input is replaced will be described. The data input from the input means is transferred to the work section A. Further, the patterns are sequentially transferred from the registered dictionary to the work part B, and it is determined whether the pattern of the work part B is included in the pattern of the work part A. If not included, the next pattern is read from the dictionary and collated again. If the collation is successful, the replacement data is read from the dictionary, and the data in the work part A is replaced, thereby terminating the process.

A more detailed flow of the above matching process will be described. FIG. 8 is a flowchart showing the processing. At the time of collation in step S31, the data of the work part A is divided. The division will be described with reference to FIG. Division is performed when an uninput portion (a portion that is not a character) exists linearly in the vertical or horizontal direction. For example,
Characters such as “Set” and “Book” are divided vertically into two parts,
A character such as "stomach" is divided into two in the horizontal direction.
For example, characters such as "rain" and "fly" cannot be divided.

In steps S32 and S33, it is determined whether the data of the work part A can be divided in the horizontal or vertical direction. Can be broken, for example, when word <br/> input to click section A data are data that can be divided in the transverse direction as shown in FIG. 9 (d) in step S37, the data in the work unit A Each divided data is copied to another work part (for example, work part C (FIG. 9)
(E)), work part D (FIG. 9 (f))... Stearyl if that could not also be divided in either the longitudinal and transverse directions
Carry out the verification and for accept the data in the work unit A in-up S34.

In step S38, the work part A and the work part B
Comparison and collation. The collation here is to compare successive parts. Here, the work part A and the work part B are divided into continuous parts. The division of the continuous part is performed as shown in FIG. The continuous part of the work part A is searched for based on the number and shape of the continuous parts included in the work part B. If a matching pattern is included in step S39, it is determined that the pattern has been matched and the replacement process is performed.

If there is no matching data, it is checked in step S40 whether or not there is a next work part (work part C). If there is, the work part is compared with the work part B (step S41), and if there is a matching pattern (step S42), it is determined that they have been collated and the replacement process is performed. If there is no matching pattern, the next work part is checked in step S43. This process is performed for all the divided workpieces. If there is no matching pattern, the process ends because no collation is performed.

Next, the flow of processing when a plurality of partial pattern data are input at the time of input will be described with reference to FIGS. The same applies to the case where partial pattern data further includes partial pattern data.

In step S51, it is checked whether or not a partial pattern is still registered in the registration dictionary (partial pattern registration dictionary 5). If there is, the pattern data is referred to in step S52. In step S53, the pattern is compared with the handwritten input pattern. In step S54, the judgment is made by the partial pattern search means 6. If the specified input pattern data is included in the handwritten input pattern, the process proceeds to step S55. Proceed with the process. If not included, the next data is read from the dictionary and the collation is repeated.

In step S55, specified partial pattern data of a character corresponding to the specified input pattern data in the dictionary is obtained, and in step S56, the specified input pattern in the input pattern data is input by the partial pattern replacing means. The pattern data is changed by replacing the data with the specified partial pattern data of the character.

The data changed in step S56 returns to step S53 and repeats the collation. As a result, even when partial pattern data at a plurality of locations is input or partial pattern data is further included in the partial pattern data, collation can be performed.

FIGS. 12 to 14 are diagrams showing how the replacement is performed using an actual example. In FIG. 12, two places are replaced with partial patterns. In the examples of FIGS. 13 and 14, the work part A and the work part B are compared first, and replacement is performed because they are collated, and the work part A of the replaced result further includes a partial pattern. Replaced,
Eventually, the character "day of the week" was obtained.

As described above, the desired data can be obtained from the output unit 8 by recognizing the finally obtained data by the ordinary handwritten character recognition unit 2 and the handwritten character recognition dictionary 3.

By using the partial pattern dictionary in reverse, it is possible to know what partial data is registered for a specific character. FIG.
Is a flowchart showing the flow of the processing.

At step S62, the character input at step S61 is converted into data.
To check if there is still data in the registration dictionary (partial pattern registration dictionary 5), and if not, output the data as it is in step S69 and end the process. If there is, collation and comparison are performed in steps S64 and S65, and replacement is performed in S67 and S68. These are the same as the replacement process described above,
Use the dictionary in reverse. Therefore, as a final output result, as shown in the flowchart, it is possible to confirm what partial data is registered for a specific character.

[0032]

According to the present invention, rather than preparing handwritten user dictionary data of one character for one character, a partial pattern having a high frequency of appearance (for example, a radical or a side) may be created. Since the user dictionary is registered, the dictionary capacity can be significantly reduced as compared with the conventional handwritten user dictionary. Further, the dictionary capacity can be reduced, so that dictionary management and the like can be easily performed.

In addition, registering a partial pattern makes it possible to apply the habit of omitting ordinary handwriting as it is, so that a comfortable input environment can be constructed.

[Brief description of the drawings]

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

FIG. 2 is a flowchart showing a flow of a partial pattern registration process.

FIG. 3 is a diagram showing a data flow of a partial pattern registration process.

FIG. 4 is a diagram showing a structure of a partial pattern dictionary.

FIG. 5 is a flowchart showing a flow of processing when an input is made from an input unit.

FIG. 6 is a flowchart showing a flow of processing when performing a collation processing with a partial pattern dictionary in the processing in claim 1;

FIG. 7 is a diagram illustrating an example of a partial pattern replacement process.

FIG. 8 is a flowchart showing a flow of detailed processing at the time of partial pattern matching.

FIG. 9 is a diagram showing division and modification of a partial pattern.

FIG. 10 is a diagram illustrating an example of a continuous part in a handwritten input character.

FIG. 11 is a flowchart showing a flow of processing according to claims 2 and 3;

FIG. 12 is a diagram showing an example of processing according to claims 2 and 3;

FIG. 13 is a diagram showing an example of processing according to claims 2 and 3;

FIG. 14 is a diagram showing an example of processing according to claims 2 and 3;

FIG. 15 is a diagram illustrating an example of a process of knowing the contents registered in the dictionary from the characters input to the input unit.

FIG. 16 is a diagram showing an example of a conventional technique of registering one user handwritten dictionary for one character.

FIG. 17 is a diagram showing an example of a conventional technique of registering one user handwritten dictionary for a character string.

[Explanation of symbols]

 Reference Signs List 1 input means 2 handwritten character recognition means 3 handwritten character recognition dictionary 4 partial pattern registration means 5 partial pattern registration dictionary 6 partial pattern search means 7 partial pattern replacement means 8 output means

Claims (2)

(57) [Claims] 1. A character processing device for comparing an input pattern input from handwriting input means such as a tablet with a character pattern registered in advance in a handwritten character recognition dictionary to obtain a recognition result for the input pattern. A partial pattern registration unit that registers an input pattern input from the input unit as a partial pattern of a specific character in the handwritten character recognition dictionary; and the input pattern is registered by the partial pattern registration unit in association with the partial pattern. A partial pattern registration dictionary that stores a pattern; a partial pattern search unit that determines whether a partial pattern registered in the partial pattern dictionary is input by handwriting in an input pattern that is input from the input unit; All or a part of the input pattern input by the Based on the results of over down search means, and a partial pattern replacement means for replacing the data read from the partial pattern registered dictionary, it has been changed by the partial pattern replacement unit pattern
To the search using partial pattern search means.
To perform the replacement by the partial pattern replacement means.
After repeated until replacement is completed, against the handwriting recognition dictionary, character processing apparatus and outputting a recognition result. 2. A method according to claim 1, wherein said partial pattern registration dictionary registers
Input pattern is the input pattern already registered in it.
2. The character processing apparatus according to claim 1, wherein registration can be performed hierarchically by using an application .