JPH10105650A - Character recognition device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a character recognition device which has a high character recognition processing speed and a high character recognition precision. SOLUTION: A first candidate output part 30 outputs the most likely candidate out of recognition candidates sent from an online recognition part 50 to a liquid crystal tablet 10 as a first candidate. Thereafter, a control part 20 sends information of the form of an entire character to an offline recognition part 70, and this part 70 extracts features of the character from the form of the entire character and refers to an offline recognition dictionary 80, where character codes corresponding to features of characters are described, to recognize the character. The candidate obtained as the result of character recognition is sent to a next candidate output part 40, and this part 40 outputs the recognition candidate obtained by the offline recognition part 70 to the liquid crystal tablet 10 as a next candidate in accordance with user's instruction.

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 As a method of 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, and a character shape is used. There is an off-line method for performing recognition by using the method.

For example, when a character "Large" is input using a pen as shown in FIG. 1A, in the online method, character recognition is performed using a time series of three strokes shown in FIG. 1B. I do. In the off-line 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 a direction component of the outline is used as a feature amount. Perform character recognition.

Conventionally, in a character recognition apparatus for inputting characters by hand using a pen and recognizing the characters in real time, an online method has been mainly used as a recognition method.

[0005]

However, in the above-described character recognition device, character recognition is performed using a time series of strokes. Therefore, if a user inputs a wrong stroke order, correct recognition cannot be performed. However, even if the characters are input in the correct writing order, there is a case where the information of the intersection of the strokes is not known, so that the stroke may be erroneously recognized.

For example, when the character "large" is input, if the input is performed in an incorrect writing order as shown in FIG. 2A, a recognition result as shown in FIG. 2B is obtained. There is a possibility that no correct answer candidate will come out.

On the other hand, if character recognition is performed by the off-line method, the recognition is performed using the shape of the character. Therefore, even if the characters are input in an incorrect writing order, there is a high possibility that correct recognition is possible. However, in the offline method, the character recognition process can proceed each time one stroke is input, whereas in the offline method, the character recognition process cannot be performed until after all the characters have been input. However, there is a problem that it takes a long time after the input of the character is completed until the recognition candidate is output.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and has as its object to provide a character recognition device which has a high character recognition processing speed and high recognition accuracy.

[0009]

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. The character recognition means for recognizing the converted character into a character code, and a display means for displaying the character recognized by the character recognition means. Online recognition means for performing character recognition using a time series of lines input by a character input means, first candidate output means for outputting a recognition candidate recognized by the online recognition means as a first candidate, And offline candidate recognizing means for performing character recognition using the form of (i) and a next candidate output means for outputting a recognition candidate recognized by the offline recognizing means as a next candidate. It is characterized in that was.

According to the character recognition device of the present invention having the above-described configuration, when the user starts inputting characters by the handwritten character input means, the stroke information is sent to the online recognition means one stroke at a time. The online recognition means performs a recognition process using the received stroke information,
The candidate output means displays, on the display means, the recognition candidate recognized by the character recognition by the online recognition means as a first candidate. afterwards,
The character input means sends information on the entire shape of the character to the off-line recognition means, and the off-line recognition means performs a character recognition process using the received character shape. When the user gives an instruction of the next candidate, the next candidate output means outputs the recognition candidate recognized by the off-line recognition means to the display means as the next candidate.

[0011] The character recognition device according to claim 2 is
The next candidate output unit integrates the recognition candidates recognized by the online recognition unit for the characters and the recognition candidates recognized by the offline recognition unit to generate the next candidates, and outputs the next candidates to the display unit.

[0012]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 3 is a block diagram showing a schematic configuration of the character recognition device according to the present embodiment. As shown in FIG. 3, 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 is connected to an online recognition unit 50 as an online recognition unit and an offline recognition unit 70 as an offline recognition unit. In addition, the control unit 20 controls the online recognition unit 5
The first candidate output unit 30 is a first candidate output unit that outputs the recognition candidate sent from the first unit to the liquid crystal tablet 10 as a first candidate, and the next candidate is a recognition candidate sent from the offline recognition unit 70. It includes a next candidate output unit 40 that is a next candidate output unit that outputs to the tablet 10. The online recognition unit 50 is connected to an online recognition dictionary 60, and the offline recognition unit 70 is connected to an offline recognition dictionary 80.

Next, the character recognition operation of the character recognition device configured as described above will be described.

The user operates the liquid crystal tablet 1 using a pen.
When input of characters is started by handwriting on the control unit 0, the control unit 2
0 sends the input stroke information to the online recognition unit 50 every time one stroke is input. The online recognition unit 50 performs character recognition while referring to an online recognition dictionary 60 that describes a character code corresponding to a time series of strokes. When the input of one character is completed, the control unit 2
0 outputs a signal indicating completion of one character to the online recognition unit 50, and the online recognition unit 50 sends a candidate for the result of character recognition to the first candidate output unit 30. The first candidate output unit 30 includes:
Of the recognition candidates sent from the online recognition unit 50,
LCD tablet 1 with the most probable candidate as the first candidate
Output to 0.

Thereafter, the control unit 20 sends the information on the shape of the entire character to the offline recognition unit 70. The offline recognition unit 70 extracts the characteristics of the character from the shape of the entire character, and extracts the character corresponding to the characteristic of the character. Character recognition is performed with reference to an offline recognition dictionary 80 in which codes are described. Then, the candidate obtained as a result of the character recognition is sent to the next candidate output unit 40, and the next candidate output unit 40 sends the recognition candidate obtained by the offline recognition unit 70 to the liquid crystal tablet 10 as the next candidate according to the instruction of the user. Output.

FIG. 4A shows a display example of the liquid crystal tablet 10. On the screen of the liquid crystal tablet 10, a text display area 100 in which recognition candidates are displayed and text is created by performing editing work such as addition and deletion of characters, and a trajectory of character input by performing character input with a pen Are provided, a recognition start button 120 for ending the input of one character and instructing the start of character recognition, and a next candidate button 130 for instructing the output of the next candidate.

Next, the operation of the present embodiment will be described by taking, as an example, a case where the user inputs the character "large" using a pen.

The user uses a pen to enter a character input area 110.
When the input of the character "large" is started, the online recognition unit 50 starts character recognition. When the user touches the recognition start button 120 with a pen at the end of one character input, the control unit 20 outputs a signal of one character end to the online recognition unit 50, and the online candidate recognizes the first recognition candidate. The input to the candidate output unit 30 and the first candidate output unit 30
Outputs the first candidate to the sentence display area 100. FIG. 4A shows an example of the result of the first candidate output when the writing order of “large” is input in the wrong order and the input is performed in the writing order of FIG. 2A. After that, since the first candidate is not the correct answer, it is assumed that the user touches the next candidate button 130 with a pen. Then
The next candidate output unit 40 displays the recognition candidates recognized by the offline recognition unit 70 in the text display area 100. FIG.
(B) shows an example in which the next candidate is output.

[0020] After starting to input the character "Large",
The operation of outputting the first candidate and the next candidate for character recognition until it is determined will be described again in detail with reference to the flowchart of FIG. First, in step 1 (hereinafter, abbreviated as S1; other steps are the same), it is checked whether or not the user is inputting a character using a pen. This can be determined by checking whether the pen is in contact with the character input area 110. If the input is being performed, in step S2, the process waits until the input of one stroke is completed. This is performed by detecting that the pen leaves the character input area 110.

Then, in S 3, stroke information (that is, a trajectory from when the pen comes into contact with the character input area 110 until it leaves) is output to the online recognition unit 50. S4
Then, the online recognition unit 50 outputs the online recognition dictionary 60
And perform online recognition processing. When the recognition process for one stroke is completed, the process returns to S1. If it is determined in step S1 that a character is not being input, it is determined in step S5 whether the recognition start button 120 has been pressed. If the recognition start button 120 has been pressed, a signal indicating the end of one character is output to the online recognition unit 50 in S6. In S5, the recognition start button 12
If 0 has not been pressed, it is checked in step S7 whether another operation (text editing operation, etc.) has been performed. If another operation has been performed, the input of one character has been completed up to that point. It moves to S6. If no other operation has been performed, S1
Return to After S6, in S8, the online character recognition unit 50
Outputs the character recognition result to the first candidate output unit 30. Then, in S9, the first candidate is output.

In the example of FIG. 4, the user indicates “large” in FIG.
Input characters in the writing order of (a) (S1, S2, S3),
Online character recognition is performed (S4), and when the user touches the recognition start button 120 with a pen (S5, S6),
The online character recognition result of FIG. 2B is input to the first candidate output unit 30 (S8). The first candidate output unit 30 outputs the most probable candidate “A” among the recognition candidates to the text display area 100 as the first candidate (S9). The display at this time is shown in FIG.

Thereafter, in step S10, the input character shape is sent to the off-line recognition unit 70,
Starts the off-line character recognition process with reference to the off-line recognition dictionary 80. Then, in S11, it is determined whether or not the next candidate button 130 has been pressed.
In step 2, the off-line character recognition result is output to the next candidate output unit 40, and the next candidate output unit 40 displays the next candidate in the text display area 100.
Output to At this time, if the first candidate output in S9 is included in the off-line character recognition candidates, the next candidate is output excluding that candidate. Then, in step S13, the user touches the desired candidate with a pen, thereby performing a process of determining the candidate and clearing the character input area 110. If the next candidate button has not been pressed in S11,
In S14, it is checked whether or not another operation (input of the next character, editing operation of a sentence, etc.) has been performed. If another operation has been performed, the off-line character recognition process is interrupted in S15. Then, in S16, it is considered that the first candidate is a desired candidate, and a determination process is performed to clear the character input area 110. If no other operation is performed in S14, S11
Return to

In the example of FIG. 4, since the first candidate "A" is not a desired candidate, if the user touches the next candidate button 130 with a pen (S11), the offline recognition unit 70 performs offline recognition. The result is input to the next candidate output unit 40. As an off-line recognition method, for example, FIG.
As shown in (C), the line of the character is thickened, the directional component of the outline is extracted as a feature of the character, and the offline recognition dictionary 80 that stores the directional component of the outline and the corresponding character code is referred to. Perform character recognition. In the off-line character recognition, character recognition is performed based on the shape of the character. Therefore, even if the stroke order is wrong, there is a high possibility that the character can be correctly recognized. In the example of FIG. 4, as a result of performing the off-line character recognition of “Large”, a candidate “Large, height, dog” is obtained. (S12). The display at this time is shown in FIG. Then, when the user touches the desired candidate "large" with a pen, "large" is determined (S13).

In the present embodiment, only the offline recognition result is displayed as the next candidate. However, the next candidate may be created by integrating the online recognition result and the offline recognition result. For example, in the example of FIG. 4, the next candidate is created by integrating the online recognition result “A, top, pa” and the offline recognition result “large, height, dog”. As an example of creating the next candidate, the offline recognition result and the online recognition result are alternately arranged and “Large,
A, height, top, dog, pa ”are generated, and the candidate“ large, height, top, dog, pa ”excluding“ A ”output as the first candidate is set as the next candidate. You may make it output.

[0026]

As is apparent from the above description, according to the character recognition device of the first aspect of the present invention, when a character is input, first, the character is recognized in an on-line system. The candidate is output, and then the character recognition is performed in an offline manner, and the next candidate is output. Therefore, even if the processing time until the first candidate is output is short and the writing order is wrong, the correct answer is given to the next candidate. Is likely to be included.

According to the character recognition device of the present invention, since both the online recognition result and the offline recognition result can be reflected in the next candidate, the recognition accuracy of the next candidate is improved. I can expect that.

[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.

FIGS. 2A and 2B are diagrams illustrating an example in which the same characters as in FIG. 1 are input in an incorrect stroke order. FIG. 2A illustrates a time series of strokes when the characters are input, and FIG. 2B illustrates online character recognition. It is a figure showing an example of a result.

FIG. 3 is a block diagram illustrating an electrical configuration of the character recognition device according to the present embodiment.

FIG. 4 is a diagram illustrating a display example of a liquid crystal tablet.

FIG. 5 is a flowchart showing an operation of character recognition.

[Explanation of symbols]

 10 LCD Tablet 30 First Candidate Output Unit 40 Next Candidate Output Unit 50 Online Recognition Unit 70 Offline Recognition Unit 100 Text Display Area 110 Character Input Area

Claims (2)

[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: a character recognition device that performs character recognition using a time series of lines input by the handwritten character input device; and First candidate output means for outputting a character-recognized recognition candidate as a first candidate, off-line recognition means for performing character recognition using the shape of the entire character, and a next candidate for the recognition candidate recognized by the off-line recognition means. And a next candidate output means for outputting the character as a character. 2. The next candidate output means integrates a recognition candidate character-recognized by the online recognition means and a recognition candidate character-recognized by the offline recognition means to generate and output a next candidate. The character recognition device according to claim 1, wherein: