JPS5975376A - On-line handwritten character recognizer - Google Patents

Abstract

PURPOSE:To simplify the constitution of a character recognizing circuit by dividing a tablet into a KANJI (Chinese character) area and an area for KANA (Japanese syllabary), marks, etc. CONSTITUTION:A tablet 1 is divided into three areas, that is, a column A for KANJI only, a column B for HIRAGANA (Japanese syllabary) only and a column C for KATAKANA (square form of Japanese syllabary) only respectively. The character information written to the tablet 1 with a pen 2 is supplied to a recognizing circuit provided to each area within a control part 3. While a dictionary is stored in a storage part 10 in correspondence to each area. In such a way, the constitution of the recognizing circuit is simplified.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field) The present invention relates to an improvement in an online handwritten character recognition device that recognizes characters written on a tablet.

(Background Art) Since the method of inputting kanji, kana, and Japanese sentences using a keyboard requires skill, an online handwritten character recognition device has recently been developed that inputs characters by writing them on a tablet with a pen. There are various methods for this recognition, but the present invention is applied to a feature extraction method that recognizes and symbolizes the characteristics of each stroke of a character, and recognizes characters based on the rank and number of symbols obtained from a symbol string based on stroke order. be done. In the feature extraction method, kanji and ``katakana°'' are mostly composed of straight lines, and their recognition methods belong to the same category, but when it comes to ``hiragana'', most of the strokes are curved lines, so the recognition method should not be changed. It won't happen.

(Problem to be solved by the invention 9) The present invention is an oscilline handwritten character recognition device that uses a pen to write a Japanese sentence containing kanji and kana on a tablet, displays the recognition result on the display, and then writes the next character on the same tablet. For example, if a tablet has an area for kanji and an area for ``hiragana,'' kanji are written in the kanji area and ``hiragana'' is written in the ``hiragana'' area according to the letter order. This simplifies the recognition method by having the tablet area determine whether the text is a hiragana character or a hiragana character.

It is also possible to have several areas for kanji, ``katakana'', ``hiragana'', alphabetic characters, Western numerals, symbols, etc.If the kanji and katakana areas are separated, the kanji ``ko5 I + 11 ko''' and 'Katakana'.

It is easy to distinguish it from other advanced machines. Characters written in each area of the tablet are guided by their own recognition circuits,
Recognized by their respective recognition algorithms.

(Structure and operation of the invention) An embodiment of the present invention will be described below with reference to the drawings.

Fig. 1 is an example of a block diagram of an online handwritten character recognition device using the conventional stroke feature extraction method, in which 1 is a tablet, 2 is a pen, 3 is a control unit, and 4, 6, and 9 are each in the control unit. A stroke recognition circuit, a character recognition circuit, and a control circuit are shown. Also, 10 is the storage unit 11, 13, 16
indicate the stroke dictionary, character dictionary, and font memory, respectively. 17 is a display device, and 18 is a printing device.

FIG. 2 is an example of a block diagram of the present invention, and the tablet 1 is as shown in FIG. A tablet with kanji and kana divided into areas, 2 is a pen, and 3 is a control unit.

The stroke recognition circuit consists of a kanji/katakana stroke recognition circuit 4 and a hiragana stroke recognition circuit 5, each of which is independently connected to the A column and B column of the tablet, and the character recognition circuit consists of a kanji character recognition circuit 6 and a hiragana character recognition circuit. Consists of a recognition circuit 7 and a katakana character recognition circuit 8.
They are connected to a kanji katakana stroke recognition circuit 4 and a hiragana stroke recognition circuit 5, respectively. Further, the 0 column of the tablet is connected to the Kanji/Katakana stroke recognition circuit 4.

10 is a storage unit, and the stroke dictionary is divided into a kanji-katakana stroke dictionary 11 and a hirakana stroke dictionary 12, which correspond to a kanji-katakana stroke recognition circuit 4 and a hiragana stroke recognition circuit 5, respectively. The character dictionary is divided into a kanji dictionary 13, a hiragana dictionary 14, and a katakana dictionary 15, which correspond to a kanji character recognition circuit 6, a hiragana character recognition circuit 7, and a katakana character recognition circuit 8, respectively.

FIG. 4 shows an example of the contents of a tablet and a pen, where 1 is a tablet and 2 is a pen. 2 is a board with an X circuit 23 printed on the top surface, 24 is a piezoelectric rubber that conducts vertically when pressed, 25 is a flexible board with an X circuit 26 printed on the bottom surface, and 27 is a board that changes color when heated and remains at room temperature. The thermal sheet returns to its original color when returned to , 28 is the heater 29 at the tip and the battery 3
It is a pen composed of 0. When you draw characters while pressing down on tablet 1 with pen 2, heater 2 will be placed on the thermal sheet.
At the same time when a character is drawn by 9, the X circuit and the X circuit are brought into conduction by pressing the piezoelectric rubber, and the position of the pen is indicated by the X and Y values.

X and Y are a thin matrix circuit. The characters drawn on the thermal sheet return to room temperature and disappear while writing in the next frame. There are tablets that use this external capacitive coupling method, electromagnetic coupling method, and ones that use electrical resistance.

FIG. 5 is an example of a kanji-katakana stroke dictionary, which is applied to characters composed of straight lines such as kanji and katakana.

Figure 6 is a part of the hiragana stroke dictionary 12. Since ``hiragana'' also has straight parts, the dictionary 11 in figure 5 is also necessary, and it may be separate from the kanji katakana stroke dictionary 11, or it may be shared. good.

FIG. 7 shows an example of the case of 10 strokes in the Kanji dictionary 13, in which JIS Kanji codes are written corresponding to stroke symbol strings, and are arranged in code string character strings that are easy to index.

FIG. 8 shows the hiragana dictionary 14.

Next, the operation will be explained.

In Fig. 1, the stroke of a character drawn with a pen 2 on a tablet 1 is detected by extracting the X and Y points according to the movement, and is sent to the stroke recognition circuit.
The direction is recognized from the position of the point, and the stroke symbol is determined by checking with a stroke dictionary as shown in FIG. When the stroke symbols of one character are determined one after another and the stroke symbol string is completed, it is checked against the Kanji dictionary 13 shown in FIG. 7, and the JIS code of that character is determined. In this method, when trying to recognize a Japanese sentence that is a mixture of kanji and hiragana or katakana, the same symbol strings appear, making it difficult to judge. For example, as mentioned earlier, the kanji ``ko''``工'' and the katakana ````゛ 工゛''
is ABA, and the symbol string for ``ki'' in hiragana is A.
It is ABB, and the kanji "Ipa" is also AABB. Also, it is difficult to distinguish between the hiragana "°'su'" and the katakana "su". In order to distinguish between these types of characters, a scribe usually makes marks that are different from the characters, or more advanced methods involve analyzing the preceding and following characters, that is, structural sentences. The former method places an excessive burden on the scribe, while the latter method requires complicated equipment. The present invention solves these drawbacks without placing too much burden on the scribe and without complicating the device.

As shown in Figure 3, the tablet shown in Figure 2 is divided into areas for Kanji, υ/ragana'', and ``Katakana''. (In the case of ゛Kanji "°'mixed"゛Japan', the grammatical part is placed in the kanji area, i.e. column A, and the ' and ゛ part' part is placed in the hiragana area, i.e. column B. Yu "baka" pass ° shall be written in the katakana area, that is, column C. The order of writing shall be in the order of the sentences as shown by the arrow. However, "ji" and "mixed" shall be written.
There is no need to leave any spaces between the lines; once the line is filled with letters, the first letter at the end of the line returns to room temperature and disappears, so you can continue writing from the beginning of the line.

(Effect of the invention) The kanji character recognition circuit 6 and the katakana character recognition circuit 8 are
The kanji katakana stroke recognition circuit 4 determines from the tablet 1 whether the stroke occurred from column A or from column C and distributes it to each circuit, making it easy to recognize kanji strokes.
It is possible to distinguish between the ``work'' of `` `` and the `` `` `` of `` '' of katakana. Also, the hiragana stroke recognition circuit 7
Since this is independent from the Kanji/Katakana stroke recognition circuit 4, straight stroke recognition and curved stroke recognition are separated, so each recognition circuit can be simplified. The present invention simplifies the subsequent recognition circuit by setting areas on the tablet according to the type of characters, and the same method can be applied to symbols and Western numerals.

[Brief explanation of drawings]

Fig. 1 is an example of a block diagram of a conventional online handwritten character recognition device, Fig. 2 is an example of a block diagram of the present invention, and Fig. 3 is an example of a block diagram of a conventional online handwritten character recognition device. Figure 4 is a diagram of the basic structure of the tablet for explanation, Figure 5 is an example of a stroke dictionary that mainly uses straight lines for kanji and katakana, and Figure 6 is for curves that are missing from Figure 5 for hiragana. An example of a stroke dictionary, Fig. 7 is an example of a kanji dictionary, and Fig. 8 is an example of a hiragana dictionary. 1...Tablet 2...Pen 3...Control unit 10...Storage unit 17...Display device 18...Printing device Patent applicant Oki Electric Industry Co., Ltd. Patent application representative Patent attorney Megumi Yamamoto -

Claims (1)

[Claims] A tablet and a pen that detect stroke order and handwriting; a storage unit that stores information on strokes forming characters; a control unit that recognizes the characteristics, order, and number of strokes and outputs corresponding characters in correspondence with the storage unit; In this online handwritten character recognition device, the tablet is divided into a kanji area and a small area (at least one ``kana'' or symbol), and a recognition section corresponding to each area is provided in the storage section and control section. This online handwritten character is characterized by simplifying the recognition of characters in the valley area by separating each character of Japanese characters including kanji and kana into a small number (at least kanji and ``kana''' or symbols, etc.). recognition device.