JPS5882373A - Online character recognizing method - Google Patents

Abstract

PURPOSE:To reduce the storage capacity of a recognition dictionary and to improve a recognition rate by fractionizing KANJI (Chinese character) which can not be recognized, recognizing the fractionized pseudo radicals, and discriminating the KANJI in the form of a set of the pseudo radicals. CONSTITUTION:Stroke information on a handwritten character is extracted on a tablet 1 and then compared by a recognition part 2 with information which shows features of partial set patterns obtained by fractionizing KANJI stored in a pseudo radical dictionary and features of patterns of the whole characters other than KANJI, thereby storing the comparison result in an input register 4. The contents of the input register 4 are compared by a selecting circuit 5 with a set of partial sets of KANJI stored in a character dictionary 6 and a character code corresponding to a character other hand KANJI, discriminating the character. Thus, KANJI is discriminated by comparison for every partial set to reduce the storage capacity required for the recognition dictionary, and a recognition rate is also improved because a KANA level, i.e. pseudo radical is used for the recognition.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a recognition method related to an online handwritten character recognition device used as an input device for information processing equipment.

Conventional handwritten character input word processors used online handwritten character recognition technology to recognize handwritten characters. There are approximately 200 characters including 26 alphabets, 46 hiragana characters, voiced marks, half-voiced marks, and 1-ro symbols, and kanji J■5C62
There are 2,965 characters in the 26 first level kanji set.

Even if kanji are used as common kanji, there are 1945 characters, totaling 2000.
It means going beyond the written word.

Now, a handwritten character input Japanese word processor that recognizes over 2000 characters in total using kanji as regular kanji has been announced, but the recognition process uses, for example, the K-, L expansion method to directly recognize kanji. There was a drawback that the amount of hardware was too large, such as front and back. In addition, handwritten kanji, not limited to the above example, are not directly recognized using the same algorithm as numbers, alphabento, hiragana, etc., so the amount of calculation for recognition is enormous and the processing time is long. In order to shorten the embedding time, this resulted in an increase in the amount of hardware (this was a major drawback from the point of view of practical application).

The purpose of the present invention is to eliminate the above-mentioned drawbacks by subdividing Chinese characters, recognizing subsets of the subdivided characters, and forming a collection of the recognized subsets (~). It is a method for identifying kanji, and it is used to reduce the number of kanji that can be recognized, etc., and also to reduce the number of kanji that need to be stored in the dictionary, such as character i'i to be recognized. Since the ``#'' can be registered in the form of a series of subsets of Mt+ characters, it has the feature of being able to reduce the increase in storage capacity due to an increase in the number of recognized characters.

Examples will be described below with reference to the drawings. 1 Figure 1 shows an embodiment of the present invention; 1 is a bronch diagram; 1 is a tablet, 2 is a recognition unit, and 3 is a subset of characters obtained by subdividing kanji and characters other than kanji (hereinafter referred to as a pseudo part). A pseudo-radical dictionary that stores the features of the neck) and pseudo-parts 1''] code, 4 is a pseudo-radical dictionary that stores the pseudo-radical codes recognized by recognition +-192, lz-4 is an input register, and 5 is an input register. A selection circuit for selecting a character from a pseudo radical code (3) of 1'1 or Nb stored in 4, and 6 a character dictionary storing pseudo radical codes and character codes. The stroke information input from the tablet notebook l-1 is sent to the recognition unit 2.

Recognition unit 2 (works to recognize each pseudo radical using the well-known stroke analysis method suitable for online handwriting, -L expansion method, etc.). ,
Use the pseudo-radical dictionary 3 that stores pseudo-radical features and pseudo-radical codes. The output of the recognition unit 2 ('f-1 pseudo radical code is output and stored each time it is recognized in the input register 4. The codes are sequentially output and input to the selection circuit 5. The selection circuit 5 selects characters based on the contents of the character dictionary 6 and the pseudo radical code output from the input register 4, and outputs the resulting character code.

FIG. 2 shows an example of the pseudo radical dictionary-3. Continuing from 000 < Hexadecimal numbers indicate pseudo radical coats, pseudo ↑lX
Each pseudo-radical is listed on the right side of the IX-do.3 In the actual dictionary, the position of each pseudo-radical is based on the recognition algorithm of recognition unit 2 (4) (feature take is written). For convenience of explanation, each pseudo-radical itself is shown here. In the pseudo-radical dictionary 3, hiragana, numbers, alpha vents, etc. other than kanji are directly stored without being subdivided into A1.

FIG. 3 shows details of the input register 4. 0 is recognition unit 2
Output from 10~(7 is 1° register 1 register in input register 4, ]8 is off 4! It'll M
, the circle indicates the output of the input register 4.

FIG. 4 shows a portion of the character dictionary 6. The third line of Figure 4 shows the pseudo radical code 176, which indicates the pseudo radical RUI-Tachi-1, and the pseudo radical code 0FI (indicated by 1, 日).
The 59-like radical indicates the character ``sound'', and the character code for the character ``sound'' is JISC622.
6 code indicates the code 3231-4
. Note that the characters inside 0 in FIG. 4 are shown for convenience of explanation, and the actual dictionary is composed of a pseudo part -i\+ code and a character code.

FIG. 5 shows the processing according to the present invention when the character "Akira" is input.
(7) The pseudo radical code input to the register 17 is shown.

Hereinafter, the online handwritten character recognition method according to the present invention will be described in detail in 1 to 1 with reference to FIG. 5, taking the character "Akira" as an example.

First of all, at the timing '■゛, the operation Nagisa from tablet 1 is ``
1" is input, "゛" is output to the recognition unit 2, and the recognition unit 2 tests whether it is registered as a pseudo radical using the pseudo part -- stationary (hereinafter referred to as dictionary) 3. However, since "゛" is not registered in Dictionary 3, the undefined code * is written as ■. Register in register 10.

Then, at timing T2, ``1/'' is input from tablet 1, and at timing T, undefined 11'' is input, along with ``1/''.
] When the recognition unit 2 tests whether the pseudo-radical ``B'' exists in the dictionary 3, the pseudo-radical ``Bi''' is registered as the pseudo-radical code 050, as shown in FIG. Write the code 050 to o register 10.

Input at timing T3 "゛" is a pseudo radical in dictionary 3
Since it is not registered in the register 11, the undefined code 1' * is registered in the register 11 while leaving the o register 10 as is.

The "'" input at timing '(4) is not registered in demultiplexing 3, but at timing T3, the pseudo radical "'" is added to the undefined "゛" and the pseudo radical "014" is added to the end 3! Since it is registered as a pseudo-radical code, the undefined code of 1 Rujista month is deleted and a new code of 014 is set.In addition, 1 pseudo-radical 1 code 050 014 is used as a new pseudo-radical. Dictionary 3 is used to test the character ``pa'' to see if it is true, but since the character ``pa'' does not exist in dictionary 3 as an independent pseudo-radical, the contents of ■o register 10 and ■rugister 11 do not change. Don't be held.

At timing ``I'' 5, it becomes clear that ``-'' is input as ``1-'' (the ``character'' is entered with pseudo radical code 004 from dictionary 3), so 00/l is entered in register 12 of 2. After that, it is checked whether pseudo-radical codes are registered in the dictionary 3 for "・'" and "1-tachi". That is, a test is performed to express all strokes in the character using the minimum number of pseudo-radical codes. In this case, "shi'" is registered as a pseudo-radical code ("
” is registered in the dictionary as a pseudo-radical code 176. Therefore, 1° register 10. I, register 11,
■2Reset register 12,■. 1 in register 10
Register 76. (Thus, the character ``tate'' indicates that it is one pseudo radical shown by the pseudo radical code 176 shown in FIG. 2.

Similarly, as shown in FIG. 5, for the character "Akira", the resulting pseudo radical code is 176, OFB.

Recognize that it is a character represented by 045.065. In addition, at the timing '■゛, the input character consisting of three pseudo radical codes is re-verified and one pseudo radical code 1 is obtained.
76, at timing T9, the character consisting of the pseudo radical code 021 and the undefined 1' and '
” to identify the character “day” with the pseudo radical code oPB,
Furthermore, the timing 'l'l l + T13 rT,
4, one pseudo radical code is identified from each input character identified as two pseudo radical codes.

In this way, pseudo-radical codes are tested using a type of longest-match method. Generally, in the case of the longest match method, a match is generally checked after all of the inputs have been input. In other words, all input strokes are judged, and if the stroke cannot be recognized as a pseudo-radical, the judgment is made excluding the last stroke, which is registered in Dictionary 3 as a pseudo-radical code. The number of pseudoradicals is a number,
Even if you add the alphabet, hiragana, symbols, etc., there are over 600 characters, and if the number is large, it will be difficult to use. It is something that exists.

As explained above, by inputting "Akira" from the tablet, the 1° registers 10 to 1 in the input register 4
3 Pseudo name in register 13 (waku code 176゜OL”B
, 045,065 are stored. The contents of the 1° register 10 to the 7 register 17 in the input register 4 are 1) sequentially derived by the conversion circuit 18 in the month of release 9 and input to the selection circuit 5. In the selection circuit 5, based on the input pseudo-radical code, the 4th south is the '1'' character - 1 ¥6
Select the character code according to JIS C6226 using .

That is, ■ in the input register 4. Register 10~■7
The pseudo radical code stored in register 17 is 176
, 01"B, 045,065, the selection circuit 5 searches the character dictionary 6, and as shown in FIG. 4, the pseudo radical code is 176, OFB, 045,065.
The character `` is indicated by the character code 3E34.
It turns out to be Akira.

As described above, the 3E34 JIS C6226 character code is outputted from the selection circuit 5, thereby recognizing that the handwritten character input from the tablet 1 is the kanji character "Ang".

As explained in detail above, in the above embodiment, a kanji is subdivided, a subset of the subdivided characters is recognized, and a kanji is created as a collection of the recognized pseudo-radicals. Show how to identify. In order to recognize pseudo-radicals here, it is sufficient to use a recognition unit 2 with a simple algorithm and a pseudo-radical dictionary 3 consisting of a very small number of pseudo-radicals compared to the number of Chinese characters. For example, the number of pseudo parts for numbers, alphabets, hiragana, symbols, and the 2965 characters in the first level kanji set of JTSC6226 is over 600, of which over 400 are for the 2965 characters in kanji. It's the neck. In this way, according to the present invention, the number of 2965 Chinese characters is reduced isometrically only during recognition, resulting in an additional effect. Furthermore, an expert in the field will understand that the function of the recognition unit 2 for recognizing the contents of the pseudo-part 3 shown in Figure 2 can be performed using a simple algorithm such as the well-known stroke analysis method. I think it's easy to understand. Furthermore, the character group book 6 for identifying characters containing kanji as a collection of pseudo-radicals is simply a pseudo-radical code and a kanji code based on J Is C6226, as shown in Figure 4. Another advantage of the present invention is that even if the number of characters including kanji is large (even if the number of characters is large, only a very small amount of memory will be increased), Since the capacity of conventional dictionaries that record the characteristics of characters including kanji is extremely small, unlike dogs, online handwritten character recognition has been developed using JiS C622, which consists of 2965 kanji characters.
Even if it is extended to the 6th level kanji set, it can be easily provided, thereby making it possible to provide an input device suitable for general information processing at a low cost.

Although the basic elements have been explained in the above embodiment, it is possible to provide a better online handwritten character recognition method by implementing various improvements as shown below, which will be explained below.

First, in the above embodiment, only one standard pseudo-radical dictionary 3 is provided, but in addition to the standard dictionary, one or more pseudo-radical correlation dictionaries for each user are provided. If the configuration is such that the pseudo radicals of characters containing pseudo radicals that are difficult to recognize can be additionally registered later, the recognition rate can be further improved.
The algorithm itself is simple and has the effect of recognizing it.

Second, in the above embodiment, the content of the character dictionary 6 was a combination of only pseudo-radical codes and character codes, but as is clear from the example in FIG. Something that consists of two letters +n, ``concubine'', ``on'', ``Chapter 1'', ``meaning'', ``-dou''(r-""-") at the top of the letter.
There are some characters such as ``Akira'' and ``Rime'' that have a ``do'' at the upper left of the character. These ``tachi'' are treated exactly the same in the above example, and there is no problem if the number of characters is about the same as in the example, but if you want to further increase the number of characters, the position information of these characters ``tachi'' can be changed to a character dictionary. 6
If A1 is included in A1, the recognition rate can be further improved. It is sufficient to use the glyph structure in item 4 of the JTSC6226 glyph index as the position information.

Thirdly, in the above embodiment, each time a new stroke is input from the tablet 1, as explained using FIG.
Looking backwards to the stroke, it is necessary to verify whether the stroke up to the above-mentioned natural stroke constitutes one pseudo radical 1 (explained).For example, the pseudo radical "3" ”, “・”, and “-” are pseudo-radical codes 050, 014, and 004, but “tate”, which has 1 in these three pseudo-radicals, is also a pseudo-radical and its code is 176. A certain thing was determined by referring to the pseudo radical dictionary 3 using the stroke of ``tate''.

However, in addition to the pseudo radical dictionary 3 as a dictionary referred to by the recognition unit 2, there are also pseudo radicals such as (pseudo radical 050) + (pseudo radical 014) + (pseudo radical 004.) - (pseudo radical 176). Providing a pseudo-radical correlation dictionary that indicates the correlation between radicals has the effect of extremely shortening the recognition processing time.

As explained in detail above, the present invention subdivides kanji that are difficult to recognize, recognizes the subdivided pseudo-radicals, and identifies kanji as a collection of pseudo-radicals. The storage capacity is extremely small, and the increase in memory capacity as the number of recognized characters increases is kept to an extremely low level.Furthermore, as explained above, the recognition rate itself can be kept high because it recognizes only pseudo radicals and kana level. A suitable online handwritten character recognition method can be provided at low cost.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention, FIG. 2 is a diagram showing an example of a pseudo-radical dictionary, FIG. 3 is a detailed diagram of an input register, FIG. 4 is a partial diagram of a character dictionary, and FIG. FIG. 5 is a diagram showing the recognition method according to the present invention. 1 Tablet, 2... Recognition unit, 3... Pseudo radical dictionary, 4... Input register, 5... Selection circuit, 6... Character dictionary, 9... Output from recognition unit 2, 10 ~17...1° register 1 in input register 4
FR register, 18...Switching circuit, 19...Output of input register 4. Patent applicant Oki Electric Industry Co., Ltd. Patent application agent Megumi Yamamoto - (15) Figure 1 Figure 3

Claims (1)

[Claims] In an online character identification method that includes a tablet that extracts stroke information of handwritten characters and a dictionary that stores character features, the information from the tablet is compared with the features of the dictionary to identify handwritten characters. a first dictionary that stores characteristics of subdivided subset patterns and the entire pattern of characters other than kanji; a second dictionary that stores character codes for collections of subsets of kanji and characters other than kanji; and a tablet. A recognition unit that compares information from the first dictionary with the output of the first dictionary, an input register that stores the recognition result, and a selection circuit that compares the contents of the input register with the second dictionary to identify characters, An online character recognition method characterized by comparing and identifying kanji characters by sound μ.