JPH0567989B2 - - Google Patents

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention includes a tablet for extracting stroke information of handwritten characters and a dictionary for storing character characteristics. This invention relates to an online character recognition method for identifying handwritten characters by comparing them with the characters.

(Prior Art) Conventionally, in this field, there is a method of character recognition as a combination of subset patterns obtained by subdividing Chinese characters, as shown in Japanese Patent Application Laid-Open No. 82373/1983.

This method, as shown in Japanese Patent Application Laid-Open No. 58-82373, subdivides kanji, recognizes subsets of the subdivided characters, and identifies kanji as a collection of the recognized subsets. This method not only reduces the number of characters to be recognized, but also reduces the storage capacity required for the dictionary, and as a result, reduces the processing time required for recognition.

For example, the character "Akira" has a partial pattern of 176 (standing),
Composed of OFB (Japanese), 045 (+), 065 (〓),
According to the explanation in JP-A-58-82373, it is identified as the character ``Akira''.

(Problem to be Solved by the Invention) However, the above method is not suitable for characters registered in the dictionary that do not have a final “ヽ”, such as “kei” or “sen” due to personal habits or misconceptions. Therefore, if you write ``〓'' or ``〓'', or if you write ``Kiyo'' and ``,'' at the end due to personal habits, that character will not be recognized because it is not registered in the dictionary.

Below, the character ``special'' is changed to ``ヽ'' like ``〓''.
Let's take the last example and explain it in detail.

FIG. 9 is a block diagram showing a conventional embodiment, in which 1 is a tablet, 2 is a stroke recognition section, and 3 is a block diagram showing a conventional embodiment.
4 is a stroke string register that stores the output of the stroke recognition unit 2; 5 is a partial pattern recognition unit that identifies partial patterns; and 6 stores partial patterns consisting of combinations of strokes. a partial pattern dictionary 7, a partial pattern detail check dictionary storing various conditions for a partial pattern, 9 a partial pattern string register, 10 a character recognition unit for identifying characters from the identified partial pattern, 11 is a character dictionary that stores characters made up of combinations of partial patterns.

Next, the operation will be explained. The strokes written on the tablet 1 are output to the stroke recognition section 2, compared with reference data in the stroke dictionary 3, recognized as a single stroke, and input into the stroke string register 4.

In this embodiment, strokes are classified into a total of 76 strokes using angles, directions, rotations, and loops. Therefore, the stroke dictionary 3 describes the characteristics of these strokes. FIG. 2 is an explanatory diagram showing strokes to be recognized. The stroke recognition method shown in FIG. 2 is not the purpose of this explanation, so its explanation will be omitted, but it can be easily recognized by the well-known stroke analysis method, K-L expansion method, or the like. SXX in the diagram (X is a hexadecimal number from 0 to 9, A to F)
indicates a stroke code, and the area to the right of the stroke code indicates its shape.

Figure 3 is an image of input when writing the character "〓". The circled numbers in Figure 3 indicate the stroke input order, and SXX indicates the input stroke code.

Figure 4 is an example of partial pattern dictionary 6, and SC
indicates a stroke code, which is shown in FIG.

No. 1 means that the partial pattern 0010 is composed of 1 strokes with stroke code SO4.

Similarly, No. 3 means that the combination of stroke codes SO2 and SO4 forms partial patterns such as outside, +, ",...".The right side of the stroke code and partial pattern code is the shape that each code represents. It is shown.

As is clear from FIG. 4, one partial pattern consists of a maximum of seven strokes. Also, the fourth
In order to avoid redundancy, only the parts necessary for this explanation have been extracted from the diagram.

Next, as mentioned above, the strokes written on the tablet 1 are recognized by the stroke recognition section 2 and input into the stroke string register 4, but the stroke example register 4 stores the strokes in the registers assigned the same number as the input order. The stroke code recognized by the recognition unit 2 is stored.

A partial pattern is recognized by the partial pattern recognition unit 5 using the stroke code stored in the stroke string register 4, the partial pattern dictionary 6, and the information in the partial pattern detail check dictionary 7.

FIG. 5 is a partial pattern recognition state diagram showing the progress of processing each time a stroke code is input. In the figure, "input order" is the same as the input order shown in FIG. 3, "input stroke" indicates its shape, and "stroke code" indicates the stroke code recognized by the stroke recognition unit 2. . As mentioned above, the stroke string register 4 has a group of registers corresponding to the "input order" in FIG. 5, and each register stores a stroke code.

FIG. 6 is an explanatory diagram showing the contents of the partial pattern detail check dictionary, and only the parts necessary for this explanation are extracted. The input order number column in FIG. 6 shows a pattern string of input order numbers to be checked.
The stroke A and B columns describe the strokes to be checked, and the check content column describes what to check. The B code column describes the partial pattern code when the candidates are narrowed down by the check. The candidate column shows the images of the candidates remaining at the time of the check. Note that this process is performed for each partial pattern on the dictionary.
The results of the check items of the target stroke are stored as reference data, and candidates are determined by matching this data with the results of the check items of the input stroke. Since it is complicated, the description method as shown in FIG. 6 is used.

Now, when SO4 "one" is input in the input order, the partial pattern 0010 becomes a candidate from the partial pattern dictionary 6 (FIG. 4).

When SO2 "|" is input in the input order, the strokes are combined with the input order, and from the stroke code 0051, 0053,
Partial patterns 0054, 0055, and 0072 are candidates.
Here, a case may be considered where the input order is a single partial pattern, but since SO2 in the input order is not registered as a single pattern, there is no partial pattern candidate for the input order.

Next, partial pattern detail check dictionary 7 (6th
) to narrow down the partial pattern candidates.
In the case of input order, since there is only one partial pattern candidate, the above check is not performed.

From No. 1 in Figure 6, the partial pattern selected when checking under the condition that "SO2 is lower than SO4" is 0051
(〓)、0054(``).Furthermore, ``The starting point of SO2 is
0051 (〓) and 0054 (") are excluded under the condition of "separated from SO4", and no partial pattern remains.

As described above, the candidates appearing from the partial pattern dictionary 6 are checked by the partial pattern detailed check dictionary 7 shown in FIG. 6, and the remaining candidates are shown in FIG.

To further explain Fig. 5, partial pattern code 1990 (〓) is a candidate as shown in Fig. 6 No. 3 for the partial pattern consisting of the input order ~, but the partial pattern with the input order as the first There is no need to think about it. This is because, as described above, in the combination of input order, anything that is not a partial pattern and becomes a character always includes the first stroke. Furthermore, since the partial patterns are defined with seven strokes or less, there is no need to consider partial patterns with more than seven strokes. Furthermore, since the strokes within one partial pattern must be written in succession, the numbers in the input order will not be skipped and a plurality of strokes will not become one partial pattern. Therefore, the lines indicating the partial patterns in FIG. 5 are continuous.

The information shown in FIG. 5 is sent to the partial pattern string register 9 and further processed by the character recognition section 10. It may be considered that the partial pattern sequence register 9 stores the state shown in FIG.

The character recognition unit 10 processes the following steps.

That is, the partial pattern string register 9 has 10 strokes, which exceeds the number of strokes required for one partial pattern to form one character (seven), so it cannot form a character by itself.

Moreover, since it becomes one character, it is expressed by a combination of some partial patterns, so the lines showing the partial patterns in FIG. 5 must be connected from the input order to form one line.

The combination of candidate partial patterns is shown in the seventh
It becomes diagram a.

In this case, the five combinations are rearranged in numerical order and checked to see if they are registered in the character dictionary 11. An example of the character dictionary 11 is shown in FIG.

FIG. 8 shows what characters are formed by the combination of partial pattern numbers and their shapes. The partial pattern column shows combinations of 4-digit partial patterns. The shape of the partial pattern is shown on the right side of the partial pattern number for reference.

The character column shows the JIS code of the character formed as a result of combining the partial patterns in the partial pattern column and the shape of the character, and the shape is shown for reference.

As described above, in the case of FIG. 5 which is currently being explained, since none of the combinations is registered in the character dictionary 11, there is no output from the character recognition unit 10.

In other words, in the above method, if you write a dot "ヽ" at the end, that character is not registered in the dictionary, so
There was a problem that the result would be erroneous recognition or reject, and if it was recognized except for the last point, erroneous recognition or reject would not occur.

Furthermore, preparing a dictionary that takes ``ヽ'' into account would result in a huge increase in dictionary capacity, which is a serious drawback from a practical standpoint.

(Means for Solving the Problems) The present invention includes a tablet for extracting stroke information of handwritten characters and a dictionary for storing characteristics of the characters, and compares information from the tablet with the characteristics of the dictionary to determine the stroke information of handwritten characters. In an online character recognition method that identifies calligraphy characters, a final stroke determination unit is provided to determine whether the XY values of the final stroke are below a certain value. If the XY value of the stroke is less than a certain value, the last stroke is removed and the character is extracted again.

(Function) According to the present invention, even if a stroke such as "ヽ" whose XY values are below a certain value is mistakenly added to the final stroke, and therefore a corresponding character candidate cannot be obtained, the final stroke determination unit Since it is determined that the XY value of the final stroke is less than a certain value, the character is extracted again excluding the final stroke, thereby obtaining the corresponding character candidate.

(Embodiment) FIG. 1 is a block diagram showing an embodiment of the present invention, in which 1 is a tablet, 2 is a stroke recognition unit, and 3 is a block diagram showing an embodiment of the present invention.
4 is a stroke string register that stores the output of the stroke recognition unit 2; 5 is a partial pattern recognition unit that identifies partial patterns; and 6 stores partial patterns consisting of combinations of strokes. a partial pattern dictionary 7, a partial pattern detail check dictionary storing various conditions for a partial pattern, 9 a partial pattern string register, 10 a character recognition unit for identifying characters from the identified partial pattern, 11 12 is a character dictionary storing characters made up of combinations of partial patterns, and 12 is a final stroke determination unit.

The present invention will be explained by taking the above-mentioned character "〓" as an example. A candidate partial pattern of the character “〓” is stored in the partial pattern string register 9,
The process in which a candidate character is not identified by the character recognition unit 10 has been described previously, and will therefore be omitted here.

The final stroke determination section will be explained below.

When a written character is identified by the character recognition unit 10, it is output as a recognition result.

However, as mentioned above, if you write down the character "〓" with "ヽ" at the end, you will not get any candidates.
The final stroke determination section 12 determines the magnitude of the final stroke. The final stroke of the character "〓" is SO4 "-", and the XY values are usually considered to be less than a certain value (for example, 3 mm). Therefore, the final stroke determining unit 12 determines that the final stroke is "small", and character recognition is performed again using the information in the partial pattern register 9 with the final stroke removed. At this time, in order to perform processing by removing the last stroke, (number of strokes) -1 is set in the register.

As mentioned above, partial patterns are stored in the partial pattern register 9 in the state shown in FIG.

The character recognition section 10 performs character recognition processing according to the contents of the register set by the final stroke determination section. That is, character candidates are considered by combining the partial patterns up to the input order shown in FIG.

The combination of candidate partial patterns is shown in the seventh
It will look like Figure b. In this case, the five combinations are rearranged in numerical order and checked to see if they are registered in the character dictionary 11 (FIG. 8). In this case, FIG. 8 shows that the partial pattern is 3443,0630, which is the character "sen".

As is clear from the above explanation, the written character is ``special'', and the character recognition unit 10 outputs a character code.

As explained above, if the final stroke is below a certain value and the written character cannot be identified, by adopting the configuration shown in Figure 9, the written character "ヽ" can be recognized. An online character recognition device can be provided.

(Effects of the Invention) As explained above, according to the present invention, it is possible to exclude the final point that humans easily make mistakes from, and to remove the final point due to human habit, thereby improving the recognition rate. There is no need to register characters in consideration of the final score in the dictionary, and the dictionary capacity can be reduced.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is an explanatory diagram of strokes to be recognized, Fig. 3 is an input image diagram when writing the character "〓", and Fig. 4 is an illustration of an input stroke when writing the character "〓".
FIG. 5 is an explanatory diagram of a partial pattern dictionary, FIG. 5 is a partial pattern recognition state diagram, FIG. 6 is an explanatory diagram of a detailed partial pattern dictionary, FIGS. 7 a and b are explanatory diagrams showing partial pattern candidates, and FIG. FIG. 9, an explanatory diagram of a character dictionary, is a block diagram showing a conventional example. 1... Tablet, 2... Stroke recognition unit,
3... Stroke dictionary, 4... Stroke string register, 5... Partial pattern recognition section, 6... Partial pattern dictionary, 7... Partial pattern detail check dictionary, 9... Partial pattern string register, 10... Character recognition Section 11...Character dictionary, 12...Final stroke determination section.

Claims (1)

[Claims] 1. Recognize strokes of handwritten characters from the first stroke to the last stroke in the input order, identify partial patterns from combinations of each recognized stroke, and combine the identified partial patterns. In addition, in an online handwritten character recognition method that identifies handwritten characters by extracting corresponding character candidates by comparing the combined partial patterns with a dictionary that stores character features, if no corresponding character candidates are extracted. When the XY value of the final stroke is a certain length, it is determined whether the XY value is a certain length or not, and if it is determined that it is less than a certain length, this final stroke is removed and the character candidate is re-extracted. An online handwritten character recognition method characterized by: