JPH06208657A - Method and device for recognizing character - Google Patents

Abstract

PURPOSE:To provide the method and device, for recognizing character capable of efficiently performing entire recognition processing although characters over a wide range are defined as recognizing objects. CONSTITUTION:When handwritten characters are inputted, a recognizing means 21 on a first stage retrieves a dictionary while narrowing down to a character group provided with the number of strokes by judging that the number of strokes for one character is coincident with the number of strokes of the character expected by an operator. When there is no intended character in the recognized result on the first stage and the continuation of recognition is instructed by a recognition continuation instructing means 22, it is possible for the inputted handwritten character to be inputted in a line character type. Then, a recognizing means 23 on a second stage recognizes the character group with the detected number of strokes from +1 to +2 as the recognizing object.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition method and apparatus, and more particularly to a character recognition method and apparatus for recognizing handwritten characters.

[0002]

2. Description of the Related Art Generally, in this type of apparatus, each stroke constituting an input handwritten character is detected. And
By extracting the feature amount of the input handwritten character based on the strokes and referring to the dictionary prepared in advance, candidate characters were searched for in the order of increasing distance, such as the first candidate, the second candidate, and so on. The user will select one of them.

[0003]

By the way, in this type of device, it is usually assumed that characters are input in a square pattern. However, in order to improve the user interface, a somewhat omitted omission (line) is acceptable. It is better to do.

In the case of regular writing, the number of strokes of a character intended to be recognized and the input stroke are the same, but when allowing line writing, they do not always match. In other words, a large dictionary search range for character recognition must be secured.

As a result, it is expected that the time required for recognition will inevitably increase. Furthermore, it can be expected that many recognition candidates will occur.

[0006]

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and a character recognition method and a character recognition method capable of efficiently performing the entire recognition process while recognizing a wide range of characters. It is intended to provide a device.

In order to solve this problem, the character recognition method of the present invention comprises the following steps. That is, in a character recognition method of receiving stroke information of a character input by handwriting, recognizing the character by referring to a predetermined dictionary, and outputting at least one candidate, the dictionary is based on the input stroke information. A first recognition step of searching within the first range for outputting a candidate group, an instruction step of instructing recognition continuation, and a first range when the recognition continuation is instructed by the instruction step. Includes a second recognition step of searching the dictionary in a different range and outputting a candidate group.

Further, the character recognition device of the present invention has the following structure.

In a character recognition device which receives stroke information of a character input by handwriting, recognizes the character by referring to a predetermined dictionary, and outputs at least one candidate, based on the input stroke information, A first recognition unit that outputs a candidate group by searching in a first range of the dictionary, an instruction unit that instructs recognition continuation, and a first range when the recognition unit gives an instruction to continue recognition. A second recognition unit that searches the dictionary in a range different from that and outputs a candidate group.

[0010]

In the process or configuration of the present invention, the first
The dictionary is searched in the range of and the recognition candidates are output. And
When the instruction to continue the recognition is given, the dictionary is searched in the second range and the recognition candidate is output.

[0011]

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

First, an outline of the character recognition device in the embodiment will be described.

In the embodiment, the number of strokes (handwriting from pen down to pen up) input is
The processing is performed assuming that the number of strokes of the character is the same, that is, that the character in the regular typeface is input. Then, the obtained first recognition candidate (or plural candidates) is displayed and selected.

When a line typeface is input, the number of strokes and the number of strokes of the intended character do not necessarily match, so in this case, the expected character is less likely to occur as a candidate.
Therefore, the operator, for example, indicates that the candidate group recognized in the first stage is not included (instructs to display another candidate). As a result, assuming that the input handwritten character may be a line typeface, the dictionary is searched extensively for the first-stage recognition process, the character recognition process is performed, and the result is displayed and selected.

By performing the above operation, the recognition processing speed at least when inputting in a regular print is improved, and the recognition processing of the line typeface is prevented. An embodiment for solving this will be described below.

FIG. 1 shows a conceptual diagram of the construction of a character recognition device in the embodiment. In the figure, 11 is a pen for inputting handwritten characters, and 12 is an input / output unit for performing handwriting input, displaying the stroke thereof in real time, and displaying a recognition result and the like. The input / output unit 12 is composed of, for example, a liquid crystal display device and a transparent coordinate input panel arranged on its display surface. Also, on the display surface,
Several frames for entering characters are displayed.
All strokes entered in one frame are judged and recognized as one character. The recognition timing is when the input is stopped and a predetermined time has elapsed, or when the input is made in the next frame. The principle of coordinate input (detection) itself does not have the characteristics of the present invention, and any method may be adopted.

Numeral 21 indicates 1 output from the input / output unit 12.
This is a first-stage character recognition means for receiving stroke data for characters and recognizing the input Slotalk number as the number of strokes of the character intended by the operator. The result (candidate group) recognized by the first-stage character recognition means 21 is sent to the input / output unit 12 and displayed.

Reference numeral 22 is a recognition continuation instructing means for detecting an instruction when it is determined that the operator does not obtain the intended recognition result by looking at the recognition result of the first stage character recognition. When this instruction is given, the stroke information that has already been input is received from the first-stage character recognition means 21 and passed to the next second-stage character recognition means 23.

The character recognition means 23 in the second stage recognizes a handwritten character on the assumption that the stroke number of the input character does not match the stroke number of the character intended by the operator.
The recognition result is output to the input / output unit 12 and displayed.

Specific operation contents will be described.

Now, suppose that the Chinese character "A" is expected and the character is input in a line typeface as indicated by reference numeral 31 in FIG. The stroke number of the Chinese character "A" is 7. However, as described above, the stroke number of the input handwritten character is processed as the stroke number of the character in the first step. recognize. As a result, candidates 33 and 34 are displayed in a window as shown by reference numeral 32 (the younger the reference numeral, the higher the priority).
In addition, in this window 32, a software-like button (hereinafter, continuation instruction button) 35 for continuing recognition as shown in the figure, and a button for canceling input characters and performing handwriting again (hereinafter, A cancel key) 36 is displayed.

Now, in the state of FIG. 3, it is understood that the operator could not recognize the character as expected, so the pen 11 touches the recognition continue button 35. As a result, based on the input stroke information, the normal character recognition processing is performed without performing the recognition processing based on the stroke number priority.

As a result, for example, the window 41 is displayed so as to overlap the window 31, as shown in FIG.
The candidate group of the character recognition result in the second stage is displayed therein. The operator sees this window 41, confirms that the intended character “A” is present, and touches the position with the pen 11 to complete the input of one character.

In the screen of FIG. 3, the meaning of the recognition continuation button 35 is the same as instructing the display of the next candidate from the operator's point of view, so that the display character may be "next candidate". Absent. Therefore, in the window 41 in FIG. 4, it may be considered that the absence of the button for continuing the recognition indicates that there is no candidate thereafter.

FIG. 5 shows a specific configuration of the character recognition device of the embodiment.

In the figure, reference numeral 100 is a CPU that controls the entire apparatus. Reference numeral 101 denotes a memory unit composed of a ROM that stores an operation processing procedure of the CPU 100 and a character font pattern, and a RAM used as a work area. In addition, in the RAM in the memory unit 101,
A buffer area corresponding to the number of input character frames is secured, and the input stroke information is stored in the corresponding buffer area.

Although the description goes back and forth, the character recognition device in the embodiment not only recognizes characters, but also executes application programs such as editing of documents based on the characters of the recognition result. However, since the present invention is not included in such processing, the description thereof will be omitted.

Reference numeral 102 denotes a display unit made of liquid crystal or the like.
A display control unit 03 controls the display of the display unit 102 under the control of the CPU 100. The display control unit 103 includes a memory (VRA) for keeping a character pattern for display (recognition result) and a trajectory of an input stroke in seconds.
M) is built in. Reference numeral 104 denotes a transparent coordinate detection plate located on the front surface of the display unit 102, and reference numeral 105 denotes a coordinate detection unit that detects a position pointed by the pen 11 and pen up / down information. Reference numeral 106 is a dictionary used for character recognition, and the number of strokes is one element as an address when referring to each dictionary. In other words, the storage position of the characteristic amount of each character is divided and managed for each number of strokes.

An I / F 107 is used to connect, for example, a printer or an external storage device. The respective constituent units described above are connected via a bus 108.

Next, the operation processing procedure of the CPU 100 in the above configuration will be described with reference to the flowchart of FIG. Naturally, the program according to this flowchart is stored in the memory unit 101.

In step S21, the stroke data for one character is received from the coordinate detecting section 105. Then, in step S22, assuming that the number of strokes in the input stroke data is the number of strokes of the input character, the search range of the recognition dictionary 106 is determined, and the data within the range and the feature amount of the input stroke information are determined. Based recognition processing is performed. In step S23, an instruction to draw an image such as a window is given to the display control unit 103, and the candidate characters obtained as a result of the recognition (correctly, the font pattern corresponding to the character code obtained by the recognition) are displayed and controlled. By transferring to the unit 103, a screen as shown in FIG. 3 is displayed.

In the next step S24, it is determined whether or not the recognition continuation button 35 has been touched. If it is determined that it has not been touched, this processing ends.

When the recognition continuation instruction button 35 is touched, that is, when it is instructed that the candidate group recognized as the input stroke number is the stroke number of the character does not have the intended one, the process is terminated. It proceeds to step S25.

In step S25, it is determined that the number of input strokes does not match the number of strokes of the target character.
Perform recognition processing.

That is, referring to the example of FIG. 3, a characteristic point is extracted from the stroke data of the character "A" intended (or expected) by the operator by a conventional algorithm, and the characteristic point is correctly 6 Although it is a stroke, it is a dictionary pattern made from a pattern written with 5 strokes. Correctly, it is 7 strokes, but by matching with the feature points of the dictionary pattern made from the pattern written with 5 strokes, Output characters with similar patterns as candidates. The output candidate group is sent to the display process in the next step S26.

As described above, according to the present embodiment, even if a character is handwritten in a line typeface, it can be recognized on the premise that it is a line typeface if the recognition continuation is instructed. Moreover, if handwritten characters are input with the same number of strokes as the number of strokes of the intended character, only a narrow range is searched in the recognition dictionary in the recognition process, so the recognition process can be speeded up and an extra range can be achieved. Since it does not search for, it is possible to provide a high-speed, high-recognition character recognition device.

[Second Embodiment] In the above-mentioned embodiment, when the stroke is input in 5 strokes in step S25, the stroke count is 6 correctly, but the dictionary pattern made from the pattern written in 5 strokes is correct. Although it has been described that 7 strokes are matched with a dictionary pattern created from a pattern written with 5 strokes, the feature points of the input pattern of 5 strokes, the dictionary pattern of 6 strokes, and the features of the dictionary pattern of 7 strokes A method of performing matching while shifting the points may be used. When input with 5 strokes, 6 strokes, 7 strokes
Although it has been described that the image (the number of strokes of the input stroke + 1 to +2) is matched with the dictionary, the recognition continuation instruction area is also provided on the candidate display screen of FIG.
It is also possible to search a dictionary of strokes and 9 strokes.

Further, the object to be recognized first is the character written with the correct stroke number, but the character written with the accurate stroke number and the character with the stroke number actually written may be the target.

[Third Embodiment] In the first embodiment, the recognition means 21 of the first stage in FIG. 1 is replaced with the character recognition means 24 for educational Chinese characters, and similarly the recognition means of the second stage is used. It is also conceivable to replace 23 with the character recognizing means 25 for the JIS first level and add a recognizing means 26 for the JIS second level.

Referring to FIG. 6, the stroke data of the handwritten character is first sent from the input / output unit 12 to the character recognition means 24 for the educational Chinese character. The character recognizing means 24 recognizes using the algorithm described in the first embodiment, but the dictionary used is for only educational Chinese characters. Educational kanji is about one-third of the JIS standard, so recognition results can be output at high speed. The results recognized here are displayed as candidates as shown in FIG. When a Kanji character other than the educational Kanji character is written when it is not a candidate for display, the recognition continuation instruction area is touched to give a recognition continuation instruction. When the recognition continuation instruction is sent from the input / output unit 12, the recognition continuation means 22 sends the stored handwritten character data to the character recognition means 25 for the JIS first level.

Reference numeral 25 is a character recognition means for recognizing JIS first-level Chinese characters excluding educational Chinese characters. The feature points of the handwritten character data sent from the recognition continuation instructing means 22 are extracted, and the candidates are output as shown in FIG. 4 by matching with the feature points of the JIS first level dictionary data excluding the educational Chinese characters. If there is no written character in the displayed recognition candidates, a recognition continuation instruction is given. When the recognition continuation instruction is issued, the recognition continuation instruction means 22 sends the stored handwritten character data to the character recognition means 26 targeted for the JIS second level.

Reference numeral 26 is a character recognition means for recognizing JIS Kanji of the second level. The feature points of the handwritten character data sent from the recognition continuation instruction means 22 are extracted, and
The candidates are output by matching the feature points of the data of the second level dictionary.

By performing the processing described above,
It is possible to provide a recognition device capable of recognizing normally used educational kanji at high speed and accurately, and also recognizing JIS second level kanji.

The present invention may be applied to a system composed of a plurality of devices or an apparatus composed of one device. Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program to a system or an apparatus.

[0045]

As described above, according to the present invention, it is possible to efficiently perform the entire recognition process while recognizing a wide range of characters.

[0046]

[Brief description of drawings]

FIG. 1 is a conceptual diagram of the configuration of a character recognition device in an embodiment.

FIG. 2 is a flowchart showing a character recognition processing procedure in the first embodiment.

FIG. 3 is a diagram showing a display example of a recognition result in the first stage of the first embodiment.

FIG. 4 is a diagram showing a display instruction of a recognition result at the second stage in the first embodiment.

FIG. 5 is a block configuration diagram of a specific character recognition device according to the embodiment.

FIG. 6 is a structural conceptual diagram of a character recognition device in a third embodiment.

[Explanation of symbols]

 11 input pen 12 input / output unit 21 first stage character recognition unit 22 recognition continuation instruction unit 23 second stage character recognition unit 100 CPU 101 memory unit 102 display unit 103 display control unit 104 coordinate detection panel 105 coordinate detection unit 106 recognition Dictionary 107 I / F 108 bus

Front page continued (72) Inventor Takashi Harada 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc. (72) Inventor Eiji Takasu 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc. (72) Inventor Katsuhiko Sakaguchi 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc.

Claims (6)

[Claims] 1. A character recognition method for receiving stroke information of a character input by handwriting, recognizing the character by referring to a predetermined dictionary, and outputting at least one candidate, based on the input stroke information. , A first recognition step of searching the first range of the dictionary and outputting a candidate group, an instruction step of instructing to continue recognition, and a first recognition step of instructing to continue recognition when the first step of recognition is instructed. And a second recognition step of outputting the candidate group by searching the dictionary in a range different from the range. 2. The first recognition step is for recognizing characters having the stroke number that matches the input stroke number, and the second recognition step is for the number of strokes that is the recognition object in the first recognition step. The character recognition method according to claim 1, wherein characters having the same number of strokes are targeted for recognition. 3. The first recognition step recognizes a character that is frequently used as a recognition target, and the second recognition step recognizes a character that is used less frequently than a character that is a recognition target in the first recognition step. The character recognition method according to claim 1, wherein the character recognition method is a target. 4. A character recognition device which receives stroke information of a character input by handwriting, recognizes the character by referring to a predetermined dictionary, and outputs at least one candidate, based on the input stroke information. First recognition means for outputting a candidate group by searching in a first range of the dictionary; instruction means for instructing to continue recognition; and if the instruction means instructs to continue recognition, the first recognition means A character recognition device, comprising: a second recognition unit that searches the dictionary in a range different from the range and outputs a candidate group. 5. The first recognizing means recognizes a character having a stroke number corresponding to the input stroke number as a recognition target, and the second recognizing means recognizes a stroke number as a recognition target by the first recognizing means. The character recognition device according to claim 1, wherein characters having the same number of strokes are recognized. 6. The first recognition means recognizes a character that is frequently used as a recognition target, and the second recognition means recognizes a character that is less frequently used than a character that is a recognition target by the first recognition means. The character recognition device according to claim 1, wherein the character recognition device is a target.