JPH10124611A - Character recognizing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a character recognizing device with high recognizing precision without damaging a recognition processing speed by operating character recognition by a character recognizing method based on a recognition mode corresponding to a character. SOLUTION: This device is provided with an on-line recognizing part 60 and an off-line recognizing part 80 as means for recognizing a handwritten character, and a recognition mode is switched according to the kind of a character by a recognition mode switching part 40 which switches those two recognition system. In a 'HIRAGANA(Japanese cursive kana character) and KATAKANA'(square Japanese syllabary)' recognition mode, an on-line recognition processing is operated for each one stroke by the on-line recognizing part 60, and a recognized candidate is displayed in a character display area. On the other hand, in a 'KANJI'(Chinese character) recognition mode, an off-line recognition processing is operated by the off-line recognizing part 80, and recognized candidate is displayed in the character display area.

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 capable of recognizing a character and converting it into a character code when a character is inputted by handwriting using a pen.

[0002]

2. Description of the Related Art Conventionally, there are two methods for recognizing handwritten characters: an online method in which one input line (hereinafter referred to as a stroke) is used as a recognition unit and recognition is performed using a time series of strokes; There is an off-line method for performing recognition using shapes.

For example, when a character "Large" is input using a pen as shown in FIG. 1A, in the online system, three strokes are input as shown in FIG. Character recognition is performed using the time series of. In the offline method, character recognition is performed by extracting some characteristic from the character shape. For example, as shown in FIG. 1C, a character line is thickened, and character recognition is performed using the directional component of the outline as a feature amount. For hiragana and katakana with a small number of strokes, the online method is superior, and in contrast, for kanji with a large number of strokes, the offline method is superior.

Conventionally, in a character recognition device that inputs characters by handwriting with a pen and recognizes the characters in real time, an online method using a time series of strokes has been mainly used as a recognition method.

[0005]

However, according to the character recognition apparatus as described above, if character recognition is performed using a time series of strokes, the entire shape cannot be captured. Will decrease. Also, if you write in the wrong writing order, you may not be able to recognize it well. Conversely, if all the off-line methods are used, the time series information of the strokes that have been input will not be used. In the online method, character recognition processing can be performed for each stroke, but in the offline method, character input processing cannot be performed until all input has been completed. Takes extra than. In addition, there is a problem that the number of strokes such as hiragana and katakana is small, and the recognition rate of characters for which it is difficult to extract features of a contour line is reduced.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problem. By performing character recognition by a character recognition method based on a recognition mode for characters, the recognition processing speed is not impaired and the recognition accuracy is improved. It is intended to provide a high character recognition device.

[0007]

In order to achieve this object, a character recognition apparatus according to a first aspect of the present invention comprises a handwritten character input unit for inputting characters by handwriting, and an input device for inputting characters by the handwritten character input unit. Character recognition means for recognizing the converted character into a character code, candidate output means for outputting a recognition candidate for the character recognized by the character recognition means, and display means for displaying the recognized character. In particular, the character recognition means includes: a recognition mode designation means for designating a recognition mode for a character when a character is inputted by the handwritten character input means; and a recognition mode designated by the recognition mode designation means. And a recognition mode switching means for switching the character recognition method based on the

According to the character recognition device of the present invention having the above configuration, when the user starts inputting characters using the handwritten character input means, the recognition mode specifying means specifies the recognition mode. Can be. When a recognition mode is designated, the recognition mode switching means switches the character recognition method based on the designated recognition mode. When the user performs character input using the handwritten character input unit, the character recognition unit performs character recognition by the recognition method switched by the recognition mode switching unit. Then, the candidate output means outputs the recognition candidates and displays them on the display means.

[0009] The character recognition device according to claim 2 is
Online recognition means functioning as the recognition method for performing recognition using the time series of input lines, and offline recognition functioning as the other recognition method for performing recognition using the shape of the entire character. Means. Therefore, if the recognition mode is designated by the recognition mode designating means, the online or offline recognition means optimal for the designated recognition mode is switched by the recognition mode switching means. Thereby, character recognition can always be performed accurately.

Further, in the character recognition apparatus according to the third aspect, the recognition mode switching means switches between the online recognition means and the offline recognition means based on a recognition mode designated by the recognition mode designation means. It is configured as follows. Therefore, for example, when the user specifies the recognition mode of “Hiragana / Katakana”,
The recognition mode is switched to the online recognition unit by the recognition mode switching unit to perform character recognition, and when the user designates a recognition mode of “Kanji”, the recognition mode is switched to the offline recognition unit by the recognition mode switching unit and the character is recognized. Recognition is performed.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a character recognition device according to the present invention will be described below with reference to the drawings.

FIG. 2 is a block diagram showing a schematic configuration of the character recognition device according to the present embodiment. As shown in FIG. 1, an input / output integrated liquid crystal tablet 10 constituting a handwritten character input unit and a display unit of the present invention is connected to a control unit 20 for controlling data transfer and the like. The control unit 20
It includes a recognition mode designating section 30 as recognition mode designating means for designating a recognition mode of a character designated by the user from the liquid crystal tablet 10. The control unit 20
Are connected to an online recognition unit 60, which is an online recognition unit of the present invention, and an offline recognition unit 80, which is an offline recognition unit of the present invention. Further, the control unit 20 controls the recognition mode switching unit 40 as a recognition mode switching unit that switches between the online recognition unit 60 and the offline recognition unit 80 according to the recognition mode data stored in the recognition mode designating unit 30. Contains. In addition, the control unit 20 includes a candidate output unit 50 as a candidate output unit that outputs a recognition candidate sent from the online recognition unit 60 or the offline recognition unit 80 to the liquid crystal tablet 10.

The online recognition unit 60 is connected to an online recognition dictionary 70, and the offline recognition unit 80 is connected to an offline recognition dictionary 90.

In the character recognition device configured as described above, when the user starts inputting characters while pressing the pen against the upper surface of the liquid crystal tablet 10 using a pen, the user may use "Hiragana / Katakana". Specify whether to recognize the character in the recognition mode of "Kanji" or in the recognition mode of "Kanji". The specified recognition mode is stored in the recognition mode specifying unit 30. The recognition mode switching unit 40 determines whether to perform character recognition using the online recognition unit 60 or character recognition using the offline recognition unit 80 in accordance with the recognition mode stored in the recognition mode designation unit 30. If the recognition is performed by the online method, the stroke information input to the online recognition unit 60 is output. If the recognition is performed by the offline method, the shape information of the character input to the offline recognition unit 80 is output. . Both the online recognizing unit 60 and the offline recognizing unit 80 output candidates of the result of character recognition to the candidate output unit
Output to 0. The candidate output unit 50 displays on the liquid crystal tablet 10 in order from the most probable candidate.

Next, a specific configuration of the present embodiment will be described.

FIG. 3 shows a display example of the liquid crystal tablet 10. On the upper surface of the liquid crystal tablet 10, a text display area 500 for displaying recognition candidates at the upper part of the screen and performing editing work such as addition and deletion of characters, and character input with a pen at the lower left of the screen are performed. A character input area 510 for displaying the trajectory of the character input, and a hiragana / katakana mode designation button 520 which is designated by the user with a pen when switching to the “hiragana / katakana” recognition mode (in the figure, “hira / kata” ), A kanji mode designation button 530 (displayed as "Kanji" in the figure) specified by the user when switching to the "kanji" recognition mode, A recognition start button 540 for instructing the start of the next candidate and a next candidate button 550 for instructing the output of the next candidate are provided.

Hereinafter, a case where the character recognition device of the present embodiment is used will be described.

Assume that the user wants to input, for example, "O". "O" uses the offline method,
It is easy to make mistakes with other candidates that look similar, such as "A".
If the processing is an online method, as shown in FIGS. 4 (a) and 4 (b), the strokes are greatly different, so that "o" and "a" will not be confused. The user says "Hiragana
By selecting the recognition mode of "Katakana", the online method is set, and "O" can be correctly input.

Next, consider a case where the user wants to input, for example, "6". “Six” becomes a stroke as shown in FIG. 5A when the online method is used. This is the same as the stroke of the “sentence” shown in FIG. 5B, and “six” and “sentence” which are not mistaken in the offline method are
It is easy to make mistakes with the online method. When the user selects the "kanji" recognition mode, the offline mode is set, and "6" can be correctly input.

Before writing a character to be input in the character input area 510, the user selects a recognition mode of "Hiragana / Katakana" or a recognition mode of "Kanji". FIG. 6 shows the flow of the processing.

First, in S500 (S indicates a step, the same applies hereinafter), the initial value of the recognition mode designation unit 30 is set to the "kanji" recognition mode. This is because kanji has more characters. Next, in S510, it is determined whether or not the Hiragana / Katakana designation button 520 has been touched with a pen in order to designate the recognition mode of “Hiragana / Katakana”. If touched (S510: Y), S53
Move to 0. In S530, the value of the recognition mode designating section is rewritten to the recognition mode of "Hiragana / Katakana", and in S5
Move to 50. In S550, the rewritten information of the recognition mode designating unit 30 is sent to the recognition mode switching unit 40, and the process ends. As a result, character recognition is performed in an online manner from the next recognition.

In S510, if the Hiragana / Katakana designation button 50 is not touched with a pen, the process proceeds to S52.
Move to 0. In S520, it is determined whether or not the kanji designation button 530 has been touched with a pen in order to designate the "kanji" recognition mode. If touched (S52
0: Y), and proceeds to S540. In S530, the value of the recognition mode designating section is rewritten to the “kanji” recognition mode,
Go to 550. If the kanji designation button has not been touched with the pen, the process ends.

The operation from when the user starts to input "O" to when the character recognition candidates are output and determined, and
The operation from the start of the input of "6" to the output of the character recognition candidates and the determination will be described in detail with reference to the flowchart of FIG.

First, in S100, it is determined whether or not the recognition mode sent to the recognition mode switching unit 40 is the "Hiragana / Katakana" recognition mode. If it is the recognition mode of “Hiragana / Katakana” (S100: Y), S11
Move to 0.

In S110, it is determined whether or not the user is inputting a character in character input area 510. If the input is being performed (S110: Y), the process waits until the input for one stroke is completed in S120. This is performed by detecting that the pen has left the character input area 510. Then, in S130, the stroke information, that is, the trajectory information from when the pen comes into contact with the character input area 510 until the pen is separated is output to the online recognition unit 60. In S140, the online recognition unit 60 sets the online recognition dictionary 70
, The online recognition processing is performed one stroke at a time.

On the other hand, when the input is not being performed in S110 (S110: N), the process proceeds to S150. In S150, it is determined whether or not the recognition start button 540 has been touched with a pen. If touched (S150: Y), S
The process proceeds to 160. Conversely, if not touched (S
150: N), since the input of one character has not been completed yet, the flow shifts to S110. In S160, a signal indicating the end of one character is sent to the online recognition unit 60. Online recognition unit 6
0, the candidate is stored in the candidate output unit 50, and in S170,
The recognition result is displayed in the character display area 500. Continue S18
At 0, it is determined whether or not the next candidate button 550 has been touched with a pen. If touched (S180: Y),
The process proceeds to S170 to sequentially display the next candidates stored in the candidate output unit 50 in the character display area 500. If not touched (S180: N), the process ends.

In S100, if it is not the recognition mode of "Hiragana / Katakana" (S100: N), ie,
If the mode is the "kanji" recognition mode, the process proceeds to S190. In S190, it is determined whether or not the user is inputting a character in character input area 510. If input is being performed (S190: Y), the process returns to S190, and S190 is repeated until input is not performed. Not being typed means
This is performed by detecting that the pen leaves the character input area 510. If not being input (S190: N),
The process moves to S200. In S200, the recognition start button 5
It is determined whether 40 has been touched with a pen. If touched (S200: Y), the process proceeds to S210. If it has not been touched (S200: N), the process proceeds to S190 because the input of one character has not yet been completed. S2
In step 10, a signal indicating the end of one character is output to the offline recognition unit 80. The off-line recognition section 80 performs off-line character recognition in S220, stores the candidates in the candidate output section 50, and shifts to S230. In S230, the recognition result is displayed in the character display area 500.

At S240, it is determined whether the next candidate button 550 has been touched with a pen. If touched (S240: Y), the process proceeds to S230, and the next candidates stored in the candidate output unit 50 are sequentially displayed in the character display area 500. If not touched (S240:
N), end the process.

When the user inputs "O", the user inputs "Hiragana.
When the recognition mode of "Katakana" is selected, S510: Y,
Steps S530 and S550 are executed. When "O" is written in the character input area 510 and the recognition start button 540 is touched, S100: Y, S110, S100
120, S130, S140, S150: Y, S160
After each step, “O” is displayed in the character display area 510 in S170. If the input character is not actually “O”, the next candidate button 550 is touched (S180: Y), and the next candidate “mu” or the like is displayed in the character display area 500 (S170).

Next, when the user selects the "kanji" recognition mode when inputting "6", S510: N, S54
0, each step of S550 is executed. Then, “6” is written in the character input area 510 and a recognition start button 540 is entered.
When is touched, S100: N, S190, S200,
After the steps S210 and S220, “Six” is displayed in the character display area 510 in S230. If the input character is actually “six”, the next candidate button 550
Is touched (S240: Y), the character display area 500
The next candidate "large" is displayed (S230).

In this embodiment, the character recognition is executed after selecting the recognition mode of "Hiragana / Katakana" and the recognition mode of "Kanji", that is, the online mode and the offline mode. Instead of 540,
A mode designation button may be touched. If the desired character is not displayed in the next candidate, the off-line method may be suitable for hiragana and katakana, and the reverse method may be used. Therefore, the method may be changed midway. One of the unavoidable problems in character recognition is that it is difficult to automatically and correctly recognize characters in combinations that are extremely likely to be misrecognized, such as katakana "e" and kanji "ko". It is sometimes difficult to see whether the character is displayed. In such a case, it is more convenient for the user to be able to specify the character type than to recognize characters in one recognition mode.
The above embodiment is effective.

[0032]

As is apparent from the above description, according to the character recognition apparatus of the first aspect of the present invention, the character recognition method is changed according to the character type, so that the character recognition can be performed with a more suitable recognition method. Can be performed, so that the recognition processing speed can be maintained and the recognition accuracy can be further improved.

Further, according to the character recognition device of the present invention, when the recognition mode is designated, the character is switched to the online recognition means or the offline recognition means which is optimal for the designated recognition mode. Character recognition can always be performed accurately according to the species.

Further, according to the character recognition device of the third aspect, the user can select the character type and perform the character recognition by the online recognition means or the offline recognition means most suitable for the character type. Therefore, the recognition accuracy can be further improved.

[Brief description of the drawings]

FIGS. 1A and 1B are diagrams showing characters input by handwriting, FIG. 1A is a diagram showing a trajectory of character input, FIG. 1B is a diagram showing a time series of strokes at the time of input, and FIG. It is a figure which shows the shape made thick.

FIG. 2 is a block diagram illustrating an electrical control configuration of the character recognition device according to the present embodiment.

FIG. 3 is a diagram illustrating a display example of a liquid crystal tablet.

FIG. 4 is a diagram illustrating a time series of strokes when handwriting is input.

FIG. 5 is a diagram showing a time series of strokes when handwriting is input.

FIG. 6 is a flowchart illustrating an operation of a recognition mode selection process.

FIG. 7 is a flowchart illustrating an operation of a character recognition process.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Liquid crystal tablet 20 Control part 30 Recognition mode designation part 40 Recognition mode switching part 60 Online control part 80 Offline control part

Claims (3)

[Claims] 1. A handwritten character input unit for inputting a character by handwriting, a character recognition unit for recognizing a character input by the handwritten character input unit and converting the character into a character code, and a character recognized by the character recognition unit A character recognition device comprising: candidate output means for outputting a recognition candidate of the character recognition means; and display means for displaying a recognized character. A character recognition apparatus comprising: a recognition mode designating unit that designates a recognition mode; and a recognition mode switching unit that switches the character recognition method based on the recognition mode designated by the recognition mode designating unit. 2. An on-line recognizing means functioning as the recognizing method for performing recognition using a time series of input lines, and the other recognizing method for performing recognition using a shape of a whole character. 2. The character recognition device according to claim 1, further comprising an off-line recognition unit that functions as a function. 3. The recognition mode switching unit according to claim 2, wherein the recognition mode switching unit switches between the online recognition unit and the offline recognition unit based on a recognition mode designated by the recognition mode designation unit. Character recognition device.