JPS6084683A - Character recognizing system - Google Patents

Abstract

PURPOSE:To suppress a classifying error rate of a basic stroke to a low level by a small processing quantity, and to obtain a high recognition rate by limiting a basic stroke group which is an object to be collated, by a recognition mode of ''Kana'' (Japanese syllabary), a Chinese character, an alphabet, etc. CONSTITUTION:Recognition mode setting information is sent to a recognition mode setting part 5, and set to a recognition mode. In a preprocessing part 4, a processing such as elimination of a noise, smoothing, etc. is executed for inputted character information, and thereafter, a stroke is cut by on-and-off information of a pen touch of an input pen 2, and stroke information of every stroke is sent to a stroke recognizing part 6. In the stroke recognizing part 6, the stroke information obtained from the preprocessing part 4 is collated to a basic stroke pattern registered in a basic stroke dictionary part 7 in accordance with the recognition mode set by the recognition mode setting part 5. In this way, the input stroke is classified to one of the basic strokes.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field) The present invention relates to a character recognition method that performs character recognition based on information on strokes that constitute a character.

(Conventional wave? 111) Conventionally, input strokes are classified into basic strokes based on information about the strokes that make up the character, and characters are recognized based on the collective force of the basic strokes. Ru. Shiriki 1 Shina Power (Ra In Japanese, kanji, hirari kuna, katakana, alphabetic characters, numbers, etc.)
Many characters are used, and there are many basic strokes for classifying the characters. However, many of the strokes that make up hiragana are not used for kanji, katakana, alphabetic characters, numbers, etc. There are also strokes that make up letters and numbers that are not used for other letters. If there are many types of characters to be recognized in this way, the number of strokes that must be prepared will also inevitably increase, so basic stroke classification errors that occur when classifying input strokes into basic strokes have caused a reduction in the recognition rate.

(Purpose) The present invention is capable of limiting basic stroke groups to be matched by recognition modes such as kana, kanji, and alphabets, keeping the classification error rate of basic strokes low with a small amount of processing, and achieving a high recognition rate. The purpose is to provide a character recognition method that can.

(Example) Hereinafter, an example of the present invention will be described in detail #11 with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a character recognition device that is an embodiment of the present invention. In the figure, 6 is a character information input device for inputting character information to be recognized, which is composed of a tablet 1 and an input pen 2. By drawing character information to be recognized on the tablet 1 using the input pen 2, You can input. 4 is a preprocessing unit that performs noise removal, normalization, etc. on the character information input from the character information input device 6; 5 is a preprocessing unit that performs noise removal, normalization, etc. on the character information input from the character information input device 6; and 5, whether the input characters to be recognized are in hiragana mode, katakana mode, alphabet mode, numeric mode, etc. 6 is a stroke recognition unit that recognizes an input cast four-way from stroke information; 7 is a recognition mode setting unit for setting whether the stroke is a typical stroke pattern used for stroke recognition; A basic stroke dictionary part, 8 is a character information processing part that processes the length of each input stroke, positional relationship, etc., 9 is a character information processing part that processes the length of each input stroke, positional relationship, etc. Character spacing tA to recognize characters? m, 10 is a character dictionary section in which multiple types of character patterns are stored, 1
Reference numeral 1 denotes an output unit that outputs the result recognized by the character recognition unit 9.

Figure 2 is an example of a basic stroke, and the stroke td
, number, representative stroke, and whether or not the strokes can constitute a character corresponding to each mode are displayed. The arrow of the representative stroke indicates the direction of pen movement. A basic stroke with rOJ in each mode string indicates that it can be a stroke that constitutes a character corresponding to that mode.

Next, the present embodiment will be described with reference to FIGS. 1 and 2.

When an operator writes characters on tablet No. 1 using input pen No. 2, the coordinate information of the pen tip of input pen No. 2 on tablet No. 1 and the pen tip of input pen No. 2 touch the tablet every 1fJI at a certain time. Information on whether or not there is a fish is sent to the preprocessing section 4. The recognition mode setting information can be set by pressing a key (not shown) provided on the character information input device 6 or by touching a predetermined area on the tablet 1 with the input pen 2. 5, and a recognition mode is set. The preprocessing unit 4 performs processing such as noise removal and smoothing on the input character information, and then turns on the pen touch of the input pen 2.

A loaf is cut out from the off information, and the stroke information for each stroke is sent to the stroke recognition section 6. Further, the length of each stroke, the start point and end point of the stroke, the coordinates of the point of change in the pen movement direction of the input pen 2, the presence or absence of intersection of each stroke, etc. are sent to the character information processing section 8. S) The o-ri recognition unit 6 compares the stroke information obtained from the preprocessing unit 4 with the basic stroke pattern registered in the basic stroke dictionary unit 7 according to the recognition mode set in the recognition mode setting unit 5. Classify the incoming cast four strokes as one of the basic strokes.

Now, assuming that the recognition mode set in the recognition mode setting section is Hiragana, the input error is the ICT number 1.2.4.6.7.8.9° in Figure 2. 10
, 11.12.13.14.15.16.17.1B,
19.20.21.22.23゜26, 27.2B,
29.30, and if the recognition mode is numeric, the input stroke is the stroke td number 1.2.5.8°12, 13.
．． 14.19.20.21.22.27.29.31.
65.65.65.65.65.65.65.65.54). The same applies to other recognition modes. As described above, when all the strokes of one character are processed by the stroke recognition unit 6, the character recognition unit 9 processes the strokes,
d IET number of each input stroke from Wh section 6,
Character information such as stroke position information, stroke intersection information, and stroke length information is obtained from the character processing information section 8, a recognition mode is obtained from the recognition mode setting section 5, and the text is opened by comparing it with a character pattern registered in the character dictionary section 10. The results are output from the output unit 11. In the previous embodiment, each recognition mode such as hiragana mode, kanji mode, katakana mode, number mode, alphabet mode, etc. was explained, but when the user sets the recognition mode, the one with fewer modes is used. The burden on people is small. Therefore, several recognition modes may be combined into one, such as a hiragana/kanji mode, a katakana alphanumeric mode, etc. In this case, in addition to the recognition mode, if the number of strokes of the input character is 7 or more, the number of strokes of the input character is ,
The basic strokes to be compared may be limited based on the number of input strokes.

(Effects) As is clear from the above explanation, according to the present invention, the basic stroke group to be matched is limited by the recognition mode, a high stroke recognition rate can be obtained with a small amount of processing, and the character recognition rate can be improved. can be increased.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a character & LiJl device which is an embodiment of the present invention, and FIG. 2 is a diagram showing basic strokes stored in the basic stroke dictionary section shown in FIG. 5 is a character information input device, 5 is a recognition mode setting section, 6 is a stroke opening section, 7 is a basic script dictionary section, 9 is a character recognition section, and 10 is a character dictionary section. Applicant Canon Co., Ltd.

Claims (1)

[Claims] A character recognition device that performs character recognition using information that traps characters in the strings that make up the characters.