JPH0258664B2 - - Google Patents

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 use online handwritten character recognition technology to recognize handwritten characters, such as numbers, alphabets,
When recognizing hiragana and kanji, there are approximately 200 characters, including 10 numbers, 26 alphanumeric characters, 46 hiragana characters, and symbols such as voiced and half-voiced marks.
Furthermore, even if we limit kanji to the JIS C6226 first level kanji set, there are 2965 characters. Even if kanji are reduced to common kanji
There are 1945 characters, totaling over 2000 characters.

Now, a handwritten character input Japanese word processor that recognizes over 2,000 characters in total using kanji as regular kanji has been announced, but its recognition processing is difficult, for example, with K--
The disadvantage was that the amount of hardware required was too large, as the L-expansion method was used to directly recognize kanji. In addition, not only the above example but also handwritten kanji are not directly recognized using the same algorithm as numbers, alphabets, hiragana, etc., so the amount of calculation for recognition is enormous and processing time is required.
In order to shorten this processing time, the amount of hardware increases, which is a serious drawback from the viewpoint of practical use.

The present invention aims to eliminate the above-mentioned drawbacks, and is aimed at subdividing kanji, recognizing subsets of the subdivided characters, identifying kanji as a collection of the recognized subsets, and identifying the kanji as a collection of the recognized subsets. This reduces the number of kanji characters to be recognized and also reduces the storage capacity required for the dictionary.Since the character dictionary to be recognized can be registered in the form of a series of subsets of the characters, the storage capacity increases as the number of recognized character systems increases. It has the characteristic of being able to keep the amount low.

Examples will be described below with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention, in which 1 is a tablet, 2 is a recognition unit, and 3 is a feature of a subset of characters (hereinafter referred to as pseudo-radicals) obtained by subdividing kanji characters and pseudo-radicals. 4 is an input register that stores the pseudo radical code recognized by the recognition unit 2; 5 is a character from one or more pseudo radical codes stored in the input register 4; A selection circuit 6 for selecting is configured like a character dictionary storing pseudo radical codes and character codes. Stroke information input from the tablet 1 is sent to the recognition section 2. The recognition unit 2 works to recognize each pseudo radical using the well-known stroke analysis method, KL expansion method, etc. suitable for online handwriting. When the recognition unit 2 performs recognition, a pseudo-radical dictionary 3 that stores pseudo-radical features and pseudo-radical codes is used. The output of the recognition unit 2 is outputted and stored in the input register 4 in the form of a pseudo-radical code each time it is recognized. The pseudo radical codes stored in the input register 4 are sequentially output and input to the selection circuit 5. The selection circuit 5 selects a character 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. The hexadecimal numbers following 000 indicate pseudo-radical codes, and each pseudo-radical is written on the right side of the pseudo-radical code. In an actual dictionary, feature data based on the recognition algorithm of the recognition unit 2 is written at the position of each pseudo-radical, but each pseudo-radical itself is shown here for convenience of explanation. Note that in the pseudo radical dictionary 3, hiragana, numbers, alphanumeric characters, etc. other than kanji are directly stored without being subdivided.

FIG. 3 shows details of the input register 4. Reference numeral 9 indicates the output from the recognition unit 2, 10 to 17 indicate the _I0 register to _I7 register in the input register 4, 18 indicates the switching circuit, and 19 indicates the output of the input register 4.

FIG. 4 shows a portion of the character dictionary 6. The third line in Figure 4 is “Tachi” indicated by the pseudo radical code 176.
The pseudo radical ``日'', which is indicated by the pseudo radical ``日'' and the pseudo radical code 0FB, indicates that it is the character ``音'', and the character code for the character ``音'' is the code 323B in the JIS C6226 code. Show that something is true. Note that the characters in parentheses in FIG. 4 are shown for convenience of explanation, and the actual dictionary is composed of pseudo radical codes and character codes.

FIG. 5 shows the pseudo radical codes input to the _I0 register 10 to _I7 register 17 of the input register 4 to show the processing according to the present invention when the character "Akira" is entered.

Hereinafter, the online handwritten character recognition method according to the present invention will be described in detail with reference to FIG. 5, taking the character "Akira" as an example. Note that the start and end of each stroke correspond to the generation and disappearance of a voltage, etc., obtained when pressing and releasing pressure on a tablet with an input pen or the like, for example.

First, in the first stroke, when the operator inputs "〓" from the tablet 1, "〓" is output to the recognition unit 2, and the recognition unit 2 uses a pseudo radical dictionary (hereinafter referred to as dictionary) 3 to generate a pseudo radical. It tests whether it is registered as a radical, but since "〓" is not registered in the dictionary 3, the undefined code * is sent to I ₀ register 10.
Register. Then, when "〓" is input from the tablet 1 in the second stroke, the recognition unit 2 tests whether the pseudo radical "〓" exists in the dictionary 3 along with the undefined "〓" at timing _T1 . Then, as shown in Figure 2, the pseudo radical code 050
Since the pseudo-radical "〓" is registered, the code 050 is stored in the _I0 register 10.

"〓" input in the third stroke is a pseudo radical that is not registered in the dictionary 3, so it is entered in the I ₀ register 1.
0 is left as is and an undefined code * is registered in the _I1 register 11.

“〓” entered in the fourth stroke is in the dictionary 3
The pseudo radical "〓" is added to Dictionary 3 in combination with the undefined "〓" which is not registered in the third stroke.
Because it is registered with the pseudo radical code 014,
I ₁ Erase the undefined code of register 11 and create a new one.
The code 014 is set. In addition, the character "〓" is used to test whether pseudo radical codes 050 and 014 are new pseudo radicals using Dictionary 3, but the character "〓" exists in Dictionary 3 as an independent pseudo radical. I ₀ register 10, I ₁
The contents of register 11 are held unchanged.

When “〓” is input in the fifth stroke, “〓”
It is determined from the dictionary 3 that the character ``is'' pseudo-radical code 004, so 004 is set in the <sub>I2</sub> register 12, and then it is checked whether pseudo-radical codes for ``〓'' and ``tate'' are registered in the dictionary 3. to test whether In other words, a test is performed to express all strokes within the character using the minimum number of pseudo-radical codes. In this case, "〓" is not registered as a pseudo-radical code, and "stand" is registered as pseudo-radical code 176 in the dictionary. Therefore, <sub>I0</sub> register 10, <sub>I0</sub> register 11, and <sub>I2</sub> register 12 are reset, and 176 is registered in <sub>I0</sub> register 10. Thus, the character ``tate'' indicates that it is a single pseudo-radical as shown in the pseudo-radical code 176 shown in FIG.

Similarly, as shown in Figure 5, for the character "Akira", the resulting pseudo radical code is 176, 0FB,
It is recognized that the characters are 045, 065. In addition, in the same way that in the 5th stroke, the input character consisting of 3 pseudo-radical codes was re-examined and 1 pseudo-radical code 176 was identified, in the 9th stroke, the character consisting of the pseudo-radical code 021 was identified. and undefined "〓" and "〓", the character "day" with pseudo radical code 0FB is identified, and furthermore, the 11th, 13th, and 14th strokes are also identified as 2 pseudo radical codes each. One pseudo radical code is identified from a character. 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, a method is used in which all input strokes are judged, excluding the last stroke that is not recognized as a pseudo-radical. Even if you add numbers, alphanumeric characters, hiragana, symbols, etc., the number is just over 600, which is not a large number.Furthermore, taking into consideration the fact that people input characters at a slow speed, we performed the test in the order in which they were entered. It is something that In addition, the end of inputting one character is done by detecting the final stroke, but specifically, the time between each stroke is monitored, and when a predetermined time is exceeded, the input operation for the next character is started. There are a method of detecting that the previous stroke was the last stroke, a method of predetermining an area for inputting one character, and a method of instructing by key operation.

As explained above, "Akira" is better than Tablet 1.
By inputting the pseudo radical code into _I0 register 10 to _I3 register 13 in input register 4,
176, 0FB, 045, 065 are stored. _I0 register 10 to _I7 register 17 in these input registers 4
The contents of are sequentially led out to the output 19 by the switching circuit 18 and input to the selection circuit 5. The selection circuit 5 selects a character code according to JIS C6226 using the character dictionary 6 shown in FIG. 4 based on the input pseudo radical code. That is, in input register 4
Since the pseudo-radical codes stored in the _I0 register 10 to _I7 register 17 are 176, 0FB, 045, 065, the selection circuit 5 searches the character dictionary 6 and finds the pseudo-radical as shown in FIG. The code is 176,
It turns out that the characters 0FB, 045, and 065 are the kanji ``Akira'', which is indicated by the character code 3E34.

As described above, the JIS C6226 character code 3E34 is output 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, numbers, alphabets, hiragana, symbols, and JIS C6626 level 1 kanji set
The number of pseudoradicals for the 2965 characters is over 600, of which over 400 are pseudoradicals for the 2965 kanji characters.
In this way, the present invention has the effect of equivalently reducing the number of 2965 Chinese characters only during recognition. Furthermore, it is easily understood by experts in the field that the function of the recognition unit 2 for recognizing the contents of the pseudo-radical dictionary 3 shown in FIG. 2 can be performed using a simple algorithm such as the well-known stroke analysis method. Think of it as possible. Furthermore, the character dictionary 6 for identifying characters including kanji as a collection of these recognized pseudo-radicals can be composed of only pseudo-radical codes and kanji codes according to JIS C6226, as shown in Figure 4, so it is difficult to identify kanji. Another advantage of the present invention is that even when the number of characters included increases, only a very small amount of memory is added, and even if the pseudo radical dictionary 3 and the character dictionary 6 are combined, the characteristics of characters including kanji can be directly recorded. Unlike the large capacity of conventional dictionaries, the capacity of online handwritten character recognition is extremely small.
Even if expanded to the C6226 first level kanji set,
It is possible to provide an input device that is easy to provide and is suitable for information processing devices in general at a low cost.

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

First, in the above embodiment, only one standard pseudo-radical dictionary 3 was provided, but in addition to the standard dictionary, one or more pseudo-radical personal dictionaries for each user were provided. If a configuration is adopted in which pseudo radicals of characters including pseudo radicals that are difficult to recognize can be additionally registered later, the recognition rate can be further improved and the recognition algorithm itself is not simple.

Second, in the above embodiment, the content of the character dictionary 6 is a combination of only pseudo-radical codes and character codes, but as is clear from the example in FIG. Those that make up a single character, those that have a 〓 at the top of the character such as ``concubine'', ``on'', ``chapter'', ``in'', and ``dong'', ``sho'', and ``rhyme''. For example, there are cases where "〓" is located 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, you can change the position information of these characters ``tachi'' to a character. If it is included in the dictionary 6, the recognition rate can be further improved. It is sufficient to use the glyph structure in item 4 of the JIS C6226 Glyph Index as position information.

Third, in the above embodiment, as explained using FIG. 5, each time a new stroke is input from the tablet 1, one pseudo radical is traced back to the first stroke up to the new stroke just written. The explanation was given to verify whether the input strokes are correct or not. For example, the pseudo-radicals "〓", "〓", and "〓" are pseudo-radical codes 050, 014, 004, but "tachi", which is a collection of these three pseudo-radicals, is also a pseudo-radical and its The code is
176 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, pseudo-radicals such as (pseudo-radical 050) + (pseudo-radical 014) + (pseudo-radical 004) = (pseudo-radical 176) Providing a pseudo-radical correlation dictionary that indicates correlations between necks 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.
First, the memory capacity required for the recognition dictionary is extremely small, and the increase in memory capacity as the number of recognized characters increases is kept extremely low.Furthermore, as explained above, what is recognized is pseudo-radicals, which are at the kana level, so the recognition rate is low. It is possible to provide a suitable online handwritten character recognition method at a low cost, and the cost itself can be kept high.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of one embodiment of the present invention, and FIG.
The figure shows an example of a pseudo-radical dictionary, Figure 3 is a detailed diagram of an input register, Figure 4 is a partial diagram of a character dictionary,
FIG. 5 is a diagram showing a recognition method according to the present invention. DESCRIPTION OF SYMBOLS 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... I in input register 4 ₀ register to I ₇ register, 18
...Switching circuit, 19...Output of input register 4.

Claims (1)

[Claims] 1. An online character recognition method for recognizing handwritten characters, comprising: a tablet for extracting stroke information of handwritten characters; and storing characteristic information of pseudo radicals including subset patterns obtained by subdividing kanji characters. A first dictionary, a second dictionary that stores character codes for pseudo-radical codes that indicate a collection of subsets of kanji, and stroke information from the tablet are compared with the contents of the first dictionary to determine pseudo-radicals. The recognition unit includes a recognition unit that performs recognition processing, an input register that stores a code indicating the recognition result of the recognition unit, and a selection circuit that compares the contents of the input register with a second dictionary to identify characters. , each time stroke information is input, it is verified whether the characteristic information of the information is registered in the first dictionary, and the pseudo radical code and/or undefined code indicating the verification result is stored in the input register. At the same time, each time the input register is stored, the content of the input register is re-verified based on the longest match method to determine whether or not the content is registered in the first dictionary, and the content of the input register is rewritten based on the verification result. An online character recognition method characterized by performing the following.