JPS61226883A - Character recognizing device - Google Patents

Abstract

PURPOSE:To simplify a correcting processing to a character recognition result by using a tablet means when an erroneous recognition or a reject occurs to a Chinese character. CONSTITUTION:In a character recognizing mode, a photoelectric converting section 10 scans on a slip and stores a character pattern on the document in a buffer memory 11. A recognition processing section 12 fetches the character pattern of the memory 11, performs a recognition processing and outputs a character recognition result to a control section 16. When moving to a character correcting mode, if an object character to be corrected is a Chinese character, a corrected character is recorded on a tablet 21. The tablet 21 outputs a coordinate data corresponding to a contact position of an input pen to a tablet/KB control section 20. The control section 20 transfers the coordinate data to the processing section 12. The processing section 12 performs the recognition processing to the transferred coordinate data, and outputs the recognition result to the control section 16. The control section 16 corrects a candidate character code to a corrected character code to display on a display section 18.

Description

[Detailed description of the invention] [Technical field of invention] The present invention relates to a character recognition device with a correction function.

[Technical background of the invention and its problems] In recent years, optical character reading devices (OCR) that can read handwritten Chinese characters have been developed. By using such Kanji OCR, it becomes possible to input character data including Kanji characters recorded on a form into a computer or the like at high speed.

By the way, in Kanji OCR, erroneous recognition and rejection of input characters may occur, so correction processing is required. Conventionally, correction processing is performed by an operator inputting correction characters from a keyboard.

However, in the conventional correction method, it is relatively simple when the recognition target is alphanumeric characters and kana characters, but in the case of kanji, it is necessary to select from candidate characters that have been converted into sounds/kun. In this case, the candidate characters will be displayed on the screen of the CRT display device. There is also a method of inputting a kanji code corresponding to a kanji for correction using a keyboard.

In this method, the operator searches for the kanji code and enters it from the keyboard.

Therefore, the conventional correction method requires troublesome correction work and has the problem of lowering the efficiency of input operations.

[Object of the Invention] The object of the present invention is to create a character that can be corrected with simple operations when a kanji is misrecognized or rejected! ! ! The objective is to provide an identification device.

[Summary of the Invention] The present invention includes a tablet input means that outputs coordinate data corresponding to a contact position of an input pen. The recognition means is a character in which the character pattern data output from the photoelectric conversion means is stored in advance! The recognition process is performed based on the standard pattern for III, and the recognition process is performed based on the pre-stored correction character pattern for the coordinate data output from the tablet input means. The transfer switching means transfers the character pattern data from the photoelectric conversion means to the recognition means in the character recognition mode, and transfers the coordinate data from the tablet input means to the recognition means in the character correction mode. The control means is configured to perform correction processing on the recognition result output from the recognition means in the character recognition mode using the recognition result output from the recognition means in the character correction mode, and output the result as a final recognition result. ing.

With such a configuration, when a kanji is misrecognized or rejected, it is possible to perform correction processing with a simple operation by using the tablet means.

[Embodiment of the Invention] An embodiment of the present invention will be described below with reference to the drawings. 1st
The figure is a block diagram showing the configuration of a character recognition device according to an embodiment. In Figure 1, the photoelectric conversion unit (scanner)
Reference numeral 10 optically scans the characters recorded on the form, converts them into character pattern data which is image data, and outputs the data. The buffer 7 memory 11 stores character pattern data (hereinafter simply referred to as character pattern) output from the photoelectric conversion section 10. The recognition processing section 12 has a changeover switch 13.
Recognition processing is performed on the character pattern output from the buffer memory 11 in response to the operation based on a standard pattern previously stored in the recognition dictionary memory 14. The recognition processing unit 12 also performs recognition processing on the coordinate data transferred from the tablet/KB control unit 20 in accordance with the operation of the changeover switch 13 based on a standard pattern stored in advance in the correction dictionary memory 15. Execute. At this time, the recognition dictionary memory 14
Alternatively, one standard pattern in the correction dictionary memory 15 is output to the recognition processing section 12.

The tablet/KB mJ 10 unit 20 performs control to transfer coordinate data output from the tablet 21 and input data input from the console keyboard (KB) 22 to the recognition processing unit 12 and the control unit 16, respectively. The tablet 21 is installed near the keyboard 22, as shown in FIG. 2, and outputs coordinate data corresponding to the contact position of the writing ben 23.

In the character recognition mode, the llllll unit 16 causes the display unit 18 to display the character recognition result output from the recognition processing unit 12 via the display control unit 17. Furthermore, in the character correction mode, the control unit 16 controls K! Characters based on the correction recognition results from the I processing unit 12! ! ! The recognition result is corrected and output as the final recognition result.

The operation of this embodiment in the character recognition IA device having such a configuration will be explained. First, letters! ! ! In the Su mode, the changeover switch 13 is operated to switch to the buffer memory 11 side by the control signal (dotted line in FIG. 1) output from the control unit 16, and the changeover switch 19 is operated to switch to the recognition word 1 memory 14 side.
Switch to side works. The photoelectric conversion unit 10 scans the form and stores character patterns corresponding to characters recorded on the form in the buffer memory 11 (normally, character patterns for one line are stored). The aX processing unit 12 takes in the character pattern stored in the buffer 7 memory 11 through the changeover switch 13 in response to a control signal from the control unit 16, and performs recognition processing based on the standard pattern stored in the recognition dictionary memory 14. Execute. The recognition processing unit 12 sends character recognition results, which are candidate characters obtained through recognition processing, to the control unit 16.
Output to. The control unit 16 stores one line of the character recognition result (candidate character code) from the zero recognition processing unit 12 in the built-in memory, and normally outputs the first candidate character as the final recognition result.

When the character correction mode is entered, the control section 16 causes the display section 18 to display the character recognition results temporarily stored in the built-in memory. That is, upon receiving the first candidate character code as a result of character recognition from the control unit 16, the display control unit 17 outputs a display character pattern corresponding to the character code to the display unit 18 for display. The operator uses the display section 18
For example, check the character a recognition results for one form displayed on the screen to check for misrecognition or rejection! ! ! Understand.

When the occurrence of misrecognition or rejection is discovered, the operator operates the cursor keys on the keyboard 22 to move the cursor on the screen of the display unit 18 to the corresponding position. This means that the character to be modified has been specified.

At this time, if the character to be modified is an alphanumeric character or a kana character, the operator inputs the modified character code from the keyboard 22. This modified character code is 8 on tablet/
The data is transferred to the control unit 16 via the control unit 20. Control part 1
Step 6 modifies the first candidate character code specified by the cursor to a modified character code and displays it on the display section 18 again. When the displayed character is confirmed by the operator, the character code is outputted from the control unit 16 as the final recognition result.

Next, if the character to be modified is a Chinese character, the operator inputs a modification command from the keyboard 22 after specifying the character to be modified with the cursor. As a result, the control section 16 operates the changeover switch 13 toward the tablet/KB control section, and operates the changeover switch 19 toward the correction dictionary memory 15 side. The operator operates the writing pen 23 to write corrected characters (kanji) on the tablet 21. When the entry is completed, a command to end the entry is input from the keyboard 22. The tablet 21 outputs a series of coordinate data corresponding to the contact position of the writing pen 23 to the tablet/KB III section 20 as described above. The tablet/KB control unit 20 controls the tablet 2
The series of coordinate data from 1 (ie, the stroke data by the writing pen 23) is transferred to the recognition processing section 12 through the changeover switch 13. Upon input of the entry end command, the control section 16 stops transferring the coordinate data to the recognition processing section 12, and causes the recognition processing section 12 to start recognition processing.

! ! ! The recognition processing unit 12 performs recognition processing on the transferred series of coordinate data based on the standard pattern stored in the correction dictionary memory 15. The correction recognition result (corrected character code) obtained by this recognition process is output to the control unit 16. The control unit 1G corrects the first candidate character code, which is the kanji specified by the cursor, into a modified character code and causes the display unit 18 to display the corrected character code again. When the displayed character is confirmed by the operator, the control unit 16 outputs the character code as the final recognition result. At this time, if the corrected characters are not appropriate, the operator inputs the next candidate command from the keyboard 22. Thereby, the control unit 16 changes the display on the display unit 18 to the corrected character code that is the next candidate for the correction recognition result. If there is no appropriate correction character in the correction recognition result, the operator will write the correction character again from the tablet 21.

In this way, the letter I! When a character recognized in the mode is incorrectly recognized or rejected, if the character is an alphanumeric character or a kana character, a correction character is input from the keyboard, and if the character is a kanji character, a correction character is input from the tablet 21. Since characters can be input simply by operating the writing pen 23 from the tablet 21 and writing the characters, even if the correction characters are Chinese characters, erroneously recognized or rejected characters can be corrected with a simple input operation. Further, even if the misrecognized or rejected characters are alphanumeric characters or kana characters, it is possible to input correction characters from the tablet 21. Therefore, even if it is impossible to input corrected characters using the keyboard 22, there is an advantage that the correcting process can be performed reliably.

[Effects of the invention] As detailed above, according to the present invention, errors in kanji can be avoided.
If a failure or rejection occurs, correction processing can be executed with a simple input operation. Therefore, the correction process for the character XKI result can be simplified, and as a result, the efficiency of the entire character recognition process can be improved.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the structure of a character recognition device according to an embodiment of the present invention, and FIG. 2 is a perspective view showing an example of the structure of a tablet and a console keyboard of the same embodiment. 10... Photoelectric conversion unit, 12... Recognition processing unit, 13.
19... Changeover switch, 14... Dictionary memory for recognition, 15... Dictionary memory for correction, 16... Control unit, 2
0...Tablet/8 control unit, 21...Tablet, 22...Console keyboard. Applicant's agent Patent attorney Takehiko Suzue Figure 1, Figure 2wA

Claims (1)

[Claims] a photoelectric conversion means that optically scans characters recorded on a form and outputs character pattern data; a tablet input means that outputs coordinate data corresponding to the position designation of an input pen; character recognition means for recognizing and processing character pattern data outputted from the tablet input means and character coordinate data outputted from the tablet input means; and character recognition means for recognizing and processing the character pattern data from the photoelectric conversion means in the character recognition mode and transferring the character pattern data from the photoelectric conversion means to the character recognition means in the character correction mode. transfer switching means for transferring coordinate data from the tablet input means to the character recognition means; and a transfer switching means for transmitting coordinate data from the character recognition means in the character recognition mode, and outputting the recognition result from the character recognition means in the character correction mode in the character recognition mode. 1. A character recognition device comprising: a control means for performing a correction process on a recognition result and outputting the result as a final recognition result.