JPH07192096A - Online handwriting recognition device - Google Patents

Abstract

(57) [Summary] [Purpose] If there is a homomorphic character in the recognized character string and it cannot be determined, the homomorphic character is automatically judged based on the recognition result information of the preceding and following character strings. [Arrangement] The homomorphic character search unit 4 searches for homomorphic characters from the results accumulated in the recognition result buffer 3. Character type acquisition unit 6
Acquires the type of front and rear characters, and the positional relationship acquisition unit 7 acquires the positional relationship with the front and rear characters. The character determination unit 8 determines the character by referring to the character type and the homomorphic character table 5, and replaces the recognition result based on the result obtained from the character determination unit 8 by the homomorphic character result operation unit 9, and the recognition result display unit 10 Display with. When there are homomorphic characters such as “force” and “mosquito”, the homomorphic character table is referred to and the homomorphic character is determined based on the preceding and succeeding characters and their positional relationship. Further, when the homomorphic characters are lined up consecutively, information for judging before and after the homomorphic character string is searched for, and all the isomorphic character strings are judged by the information.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an on-line handwritten character recognition device, and more specifically, to an isomorphic character that cannot be identified as a single character such as "force" or "ka", other than a character string. The present invention relates to an on-line handwritten character recognition device for identifying characters without using linguistic information based on the recognition result of the characters.

[0002]

2. Description of the Related Art As a method of recognizing a graphic entered on a tablet, an input graphic stroke is converted from a shape feature into a set of line segment element sequence data by arcs or straight lines, and is previously converted into similar line segment element sequence data. The recognition was performed by matching the defined registered figure. There is also a method that introduces ambiguity when converting to line segment element sequence data, but it requires registration of figures. Further, there is a method of examining the distribution of the angle histogram in the writing direction, determining the line segment based on the distribution situation, and identifying the number of sides or the like. In either method, uniform processing was performed regardless of the size of the input figure and the length of the stroke. Further, in the conventional online character recognition device for inputting characters in a frame, "force" and "mosquito" are identified by the position and size of the characters written in the frame.

[0003]

However, in the conventional on-line handwritten character recognition device, it is a heavy burden for the user to perform the above transcription, and in the case of ambiguous writing, which one is to be judged? I don't know if it will be done. Further, in the case where there is no input frame, the above method cannot be taken, and it is impossible to write "power" or "power".

The present invention has been made in view of such a situation, and basically, when there is a homomorphic character in a recognized character string and it is not possible to make a judgment, basically the information on the recognition result of the preceding and following character strings is used. It is an object of the present invention to provide an on-line handwritten character recognition device that automatically determines the homomorphic character.

[0005]

In order to achieve the above object, the present invention provides (1) an on-line handwritten character recognition device for collating characters based on position information of a pen sent from a tablet. A homomorphic character table having, and a discrimination means for discriminating the homomorphic character based on the preceding and succeeding characters and the positional relationship between the preceding and succeeding characters when the homomorphic character exists, and (2) When the homomorphic characters are continuously arranged, a search unit that searches for information to be discriminated before and after the homomorphic character string and all the isomorphic character strings are discriminated by the information retrieved by the search unit. Or (3) a character recognition means for recognizing a character by the data sent from the input means, and a result recognized by the character recognition means. A recognition result buffer to be stored, a homomorphic character table having homomorphic characters and information about the same, an isomorphic character searching means for searching a homomorphic character from the homomorphic character table and the contents of the recognition result buffer, and an isomorphic character searching means. A character type acquisition unit that acquires the type of characters before and after the character that is made to be a character is provided, and the same-shaped character is determined, and further, (4) the positional relationship that acquires the positional relationship with the characters before and after the character. It is characterized in that an acquisition means is provided.

[0006]

The character is recognized by the data sent from the input means, and the recognized result is stored in the recognition result buffer.
A homomorphic character is searched from the information of the homomorphic character table having the homomorphic character and information about the homomorphic character and the information of the recognition result buffer, and the types of characters before and after the homomorphic character are acquired. If it cannot be determined based on the type of characters before and after, the positional relationship with the characters before and after is acquired, and the homomorphic characters are determined.

[0007]

Embodiments will be described below with reference to the drawings. FIG. 1 is a block diagram for explaining an embodiment of an online character recognition device according to the present invention. In the figure, 1 is a tablet, 2 is a recognition unit, 3 is a recognition result buffer, 4 is a homomorphic character search unit, 5 Is a homomorphic character table, 6 is a character type determination unit (character type acquisition unit), 7 is a positional relationship acquisition unit, 8 is a character determination unit, 9 is an isomorphic character result operation unit, and 10 is a recognition result display unit.

The character recognition unit 2 recognizes characters based on the data sent from the tablet 1. Recognition result buffer 3
Stores the recognized result. The homomorphic character search unit 4 searches for homomorphic characters from the recognized result. The homomorphic character table 5 stores homomorphic characters and information about them.
The character type acquisition unit 6 acquires the type of characters before and after. The positional relationship acquisition unit 7 acquires the positional relationship with the preceding and succeeding characters. The character determination unit 8 determines a character by referring to the character type and the isomorphic character table 5. The isomorphic character result operation unit 9 is the character determination unit 8.
Swap the recognition results based on the results obtained from. The recognition result display unit 10 displays the recognition result buffer 3 after being operated by the isomorphic character result operation unit 9.

The operation of the online handwritten character recognition apparatus according to the present invention will be described below based on a more specific example. The data sent from the tablet 1 is character-recognized by the character recognition unit 2 and stored in the recognition result buffer 3. When the recognition end signal is sent from the character recognition unit 2, the homomorphic character search unit 4 searches the homomorphic character table 5 and the contents of the recognition result buffer 3 to find out which recognition result has the same character.

For example, if there is a recognition result as shown in FIG. 2 (b) for the input data shown in FIG. 2 (a) and there is a table as shown in FIG. The search unit 4 searches the i-th table and determines that the result of the n-th character has homomorphic characters. The i-th homomorphic character is as follows.・ At least one of "Ka" is katakana either before or after.・ "K" is a number in the previous character.・ "Power" is other than the above. Is described.

In the character type acquisition unit 6, the recognition result buffer 3
Further, the type of characters before and after the character that is regarded as the homomorphic character by the homomorphic character search unit 4 is acquired. If the front and rear are homomorphic characters, it is undefined. The character determination unit 8 refers to the preceding and following character types and the isomorphic character table 5, and processes isomorphic characters according to the rules of the isomorphic character table 5.

Normally, it is difficult to judge the kanji "1" and "-(long sound)" by themselves only by handwriting. So long sound is "the previous character is katakana"
By using this characteristic, it can be identified by making a long sound only when the type of the previous character is katakana. When the homomorphic characters are lined up consecutively, the character type is undefined because the front and rear are homomorphic characters. In that case, since it cannot be determined, a search is performed until the characters are not homomorphic, and the character type is used to make the determination.

For example, in the case of the recognition result as shown in FIG. 2 (b), since the preceding character is "no" and is hiragana, and the following character is "even" and kanji, the rule of the table is " The result is "power". However, "evening"
The character is also registered as an isomorphic character as shown in FIG. 3, and it cannot be determined as a kanji at this point. Similarly,
Since the next "mouth" is also an isomorphic character, the character type Katakana can be obtained at the time of the character "G". When Katakana can be obtained with this "Gu", by giving information that the latter character is Katakana to a series of isomorphic character strings, "Power" becomes "Ka", "Evening" becomes "Ta", " The mouth can be confirmed as "b". Therefore, the homomorphic character result operation unit 9 operates the result in the recognition result buffer 3 based on the above result and sends it to the recognition result display unit 10. In addition, in FIG. 3, the place where there is "none" is that it does not depend on whatever it is.

FIG. 5 is a flow chart for explaining the operation of the online character recognition device according to the present invention. Hereinafter, each step will be described in order. First, a character string is acquired based on the recognition result (step 1), and one character is acquired from the character string (step 2). Next, it is determined whether the acquired one character is the last character (step 3), and if it is the last character, it is output (step 11). If the acquired one character is not the last character, then it is judged whether or not there is a character having the isomorphic character (step 4). If not, the process proceeds to the step 2.

If there is a character having a homomorphic character, the character types before and after are acquired (step 5), and it is determined whether the homomorphic character continues (step 6). If the homomorphic characters continue, a search is performed until the homomorphic characters do not continue (step 7), and the process goes to step 5. If homomorphic characters do not continue in step 6, the positional relationship with the preceding and succeeding characters is acquired (step 8), and it is determined whether they match the table (step 9). If they match, the results are exchanged (step 10), and the process goes to step 2, and if they do not match, the process immediately goes to step 2.

That is, in the case as shown in FIGS. 2 (a) and 2 (b), since "no" has no isomorphic character, the process proceeds to the next character. Since there is a homomorphic character in "Power", the process moves to processing of the homomorphic character. In step 5, the preceding and following character types are acquired, but in step 6, the rear part is judged to be "evening" and the character is a homomorphic character, so in step 7, a search is made up to the "mouth" where the isomorphic character does not continue. Then, the character behind is "G" and it is understood that it is a katakana character type, and since it matches the table of "K" in step 10, the result is replaced with "K".

When the character type is set to katakana as shown in the table at the same time as "ka" is replaced, the result can be replaced with "ta" for "even" and "ro" for "mouth". Like On the other hand, in the case of "Kiki Ippai", since "Kai" is Kanji,
"Power" and "Ichi" are recognized as kanji.

In the case of "," and "'", it is not possible to judge by the type of characters before and after, so the description as shown in FIG. 3 is given. In step 8, the positional relationship with the preceding and succeeding characters is acquired, which is determined by the character determination unit 8 from FIG. 3, and is sent to the isomorphic character result operation unit 9 for determination. In the case as shown in FIG. 4, since the character is lower than the previous character, the positional relationship acquisition unit 7 obtains the information “below” and the character determination unit 8 determines “,”. .

[0019]

As is apparent from the above description, according to the present invention, in an online handwritten character recognition device for collating characters based on the position information of the pen sent from the tablet, "force" and "mosquito" etc. When there are homomorphic characters, the homomorphic character table is referred to, the homomorphic character has a judgment means for discriminating it based on the front and back characters and the positional relationship with it, and when the homomorphic characters are lined up continuously, Since it has the above-mentioned means for searching the information to determine before and after the homomorphic character string and discriminating all the isomorphic character strings based on that information, a simple table can be created without using a large system such as word matching. Just by doing, it is possible to determine the homomorphic characters.

[Brief description of drawings]

FIG. 1 is a configuration diagram for explaining an embodiment of an online handwritten character recognition device according to the present invention.

FIG. 2 is a diagram showing an example of a character recognition result according to the present invention.

FIG. 3 is a diagram showing an example of an isomorphic table according to the present invention.

FIG. 4 is a diagram showing an example of position determination in the present invention.

FIG. 5 is a flowchart for explaining the operation of the online handwritten character recognition device according to the present invention.

[Explanation of symbols]

1 ... Tablet, 2 ... Recognition unit, 3 ... Recognition result buffer,
4 ... Isomorphic character search unit, 5 ... Isomorphic character table, 6 ... Character type determination unit (character type acquisition unit), 7 ... Positional relationship acquisition unit, 8 ...
Character determination unit, 9 ... Isomorphic character result operation unit, 10 ... Recognition result display unit.

Claims (4)

[Claims] 1. An online handwritten character recognition device for collating characters based on position information of a pen sent from a tablet, wherein an isomorphic character table having isomorphic character information and the isomorphic character table when the isomorphic character exists. An online handwritten character recognizing device, comprising: a reference unit for determining the same-shape character based on a front-back character and a positional relationship between the front-back character and the front-back character. 2. When the homomorphic characters are continuously arranged, a retrieval unit that retrieves information to be discriminated before and after the homomorphic character string, and all the isomorphic character strings are discriminated by the information retrieved by the retrieval unit. The on-line handwritten character recognition device according to claim 1, further comprising a discrimination means. 3. A character recognition means for recognizing a character by the data sent from the input means, a recognition result buffer for storing a result recognized by the character recognition means, a homomorphic character table having homomorphic characters and information about the same. And an isomorphic character search means for searching for an isomorphic character from the isomorphic character table and the contents of the recognition result buffer, and a character type acquisition means for acquiring the type of a character before and after the character which is an isomorphic character by the isomorphic character search means. An on-line handwritten character recognizing device, characterized by comprising: 4. The on-line handwritten character recognition device according to claim 3, further comprising a positional relationship acquisition means for acquiring a positional relationship with the preceding and succeeding characters.