JPS63276961A - Character recognition system using facsimile equipment - Google Patents

Abstract

PURPOSE:To improve the accuracy of the result of recognition by decoding a coded picture signal of a sheet sent by a facsimile equipment, detecting the position of a control mark in the sheet and measuring the length of the control mark. CONSTITUTION:When an interface circuit 10 receives a facsimile signal, an MH coded picture signal is extracted and sent to a decoding circuit 11. The circuit 11 sends a decoded original picture signal to a character line detection circuit 12. The circuit 12 uses the sent original picture signal to detect upper end scanning lines LT, RT in the subscanning direction of character line designation marks 2, 3 and lower end scanning lines LB, RB to measure the distance between the scanning lines LT and RB and between the lines RT and RB. When the measured distance is within a permissible value, a character recognition circuit 13 detects the character position and recognizes it depending on the original picture signal and the location of the marks 2, 3 in the main scanning direction. The interface circuit 14 sends the recognized result externally. Thus, the accuracy of the result of recognition is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION (1) Description of the field to which the invention pertains The present invention relates to a character recognition method using a facsimile that recognizes character information in sheet information transmitted by facsimile.

(2) Description of conventional technology Conventionally, in the recognition of character information in sheet information transferred by encoded facsimile signals, the number of scanning lines that have received code errors within one page or within a specified number of scanning lines is calculated as M number. However, if this value exceeds a specified value, the recognition process is stopped as an abnormality in the image signal. For this reason.

Even if the scanning line with the coding error does not exist in the text area, or even if it exists in the text area, it does not affect recognition, the image will be deleted if the total number of errors within one page or within the specified number of scanning lines exceeds the allowable value. Recognition processing has been canceled due to a signal abnormality. therefore,
The sender of the sheet may not be processed even though there is no code error in the character part, or the sheet may be processed even though the total number of code errors in the nine character part is small but there is a high possibility of misrecognition.

(3) Purpose of the invention The purpose of the present invention is to accurately read a sheet from a facsimile machine, and even if there is a code error in the sheet information, if it is not a character part, the image signal will not be considered abnormal and will be processed as a normal image signal. Even if there is a code error in the character section, if the size of the character line is outside the allowable value, it will be processed normally, and if it is not, it will be recognized as an abnormality in the image signal and will not be processed. To provide a means for accurately determining the normality of character information transmitted from a remote location.

(4) Structure of the invention (4-1) Features of the invention and differences from the prior art
By checking the length of the character line in the sub-scanning direction, even if a code error occurs and the original image signal disappears, it is possible to detect code errors other than character lines by determining whether the number of deleted scanning lines is within the allowable value. The main feature is that it can also be determined whether the image signal corresponding to one character line is acceptable and normal. This method differs greatly from conventional technology in terms of the accuracy of recognition results.

(4-2) Example FIG. 2 is a diagram showing an example of a sheet used in the present invention, where l is a sheet, 2 is a character line designation mark on the left, 3 is a character line designation mark on the right, 4 is a character entry frame, and the 9-character line designation mark 2.3 is black, and the 1-character entry frame 4 is located at a predetermined position from the character line designation mark 2.3, and is written in a dropout color.

Figure 3 (A), (B), and (C) are Modifi
FIG. 3 is a diagram showing an example of a decoding processing method of an edHuff+an (MH) code, where nl-n5 is the original image signal of each decoded scanning line of the 0th order. Code error processing will be explained using FIG. 3. First, it is assumed that when the MH code is normally decoded, an original image signal as shown in FIG. 3(A) is obtained. If there is an error in the run-length code corresponding to scanning line n3 in the MH code, the original image signal of scanning line n3 is changed to n2.
(previous line replacement), resulting in an original image signal as shown in FIG. 3(B). Also, En indicating the end of scanning line n3
d ofLine (EOL) If there is an error in the code, scan line n3. n4 is replaced with the original image signal of scanning line n2 (line missing), resulting in an original image signal as shown in FIG. 3(C). Generally, code errors are handled in this manner, so that when the code error shown in FIG. 3(B) occurs at the beginning of a character line, or as shown in FIG. 3(C).

If a code error exists, the number of scanning lines in a character line is reduced. Furthermore, if a code error as shown in FIG. 3B occurs immediately after the end of a nine-character line, the scanning line of the character line will be detected to be long.

FIG. 4 is a diagram showing an example of a character line detection method.
is the scanning line at the top of character line designation mark 2.

RT is the scanning line at the top of the character line designation mark 3, LB is the scanning line at the bottom of the character line designation mark 2, and RB is the scanning line at the bottom of the character line designation mark 3, which is the 0th order.

The character line detection method will be explained with reference to FIG. However, it is assumed that Sheet 1 is input by facsimile with the upper left facing upward, and the 9-character line designation mark 2.3 is read with a width of W pixels. First, character line designation marks 2 and 3 on each scanning line.
Examine the original image signal near where . Then, the upper end of the character line designation mark 2 is determined by first detecting the scanning line LT in which there is a black pixel of W pixels continuous at the left end, and f&
Scanning line RT where there is a black pixel with continuous W pixels at the right end for the first time
4. Determine the upper end of the character line designation mark 3 by detecting 0. Next, scan line LB with a black pixel continuous for W pixels at the left end.
The lower end of character line designation mark 2 is determined by detecting that there is no black pixel that is continuous for W pixels on the next scanning line, and the scanning line RB that has a black pixel that is continuous for W pixels at the right end and the next scanning line are The lower end of the character line designation mark 3 is determined by detecting that there are no black pixels that are continuous by W pixels. However, the value of the W pixel is determined within a range that takes into consideration the line density of the facsimile, the width of the character line designation mark 2°3, and the reading error of the facsimile.

FIG. 1 is a block diagram of an embodiment of the present invention.

lO is an interface circuit that performs facsimile procedures;
11 is a decoding circuit that decodes the MH code, 12 is a decoding circuit 1
The character line designation mark is determined from the original image signal sent from 1.
A character line detection circuit that checks the sub-scanning length of No. 3 and determines whether the character line is normal.

13 is a character recognition circuit that recognizes the original image signal of a character line sent from the character line detection circuit 12; 14 is a character recognition circuit r
This is an interface circuit for transmitting the recognition results sent from the character line detection circuit 3 and the image signal abnormality notification from the character line detection circuit 12 to an information processing device or the like.

Next, FIG. 1 will be explained. First, interface circuit 1
When 0 receives the facsimile signal, it extracts the MH encoded image signal and transmits it to the decoding circuit 11.

The decoding circuit 11 decodes the received MH encoded image signal and transmits the original image signal to the character line detection circuit 12.

However, when a code error is detected, the 0 character line detection circuit 12, which performs the decoding process using the method explained in FIG. The lines LT and RT and the lower end scanning line LBRB are detected, and the distance between the scanning lines LT and LB and the distance between the scanning lines wART and RB are checked. and.

If the distance between the scanning lines LT and LB and the distance between the scanning lines RT and RB are within the allowable values, the original image signal between the scanning lines LT and RB and the main scanning position of the character line designation mark 2°3 are converted into characters. It is transmitted to the recognition circuit 13.

If the value is other than the allowable value, the interface circuit 14 is notified that the character line cannot be subjected to LWm. however.

The values of the distance between scanning line '3 HLT and LB and the distance between scanning line RT and RB are determined in consideration of the case where the character line designation mark 2.3 is input by fax at an angle, and the reading error during facsimile input.0 The character recognition circuit 13 receives the original image signal received from the character line detection circuit 12 and the character line designation mark 2.
．． The character position is detected and recognized based on the position in the main scanning direction of 3. Then, the recognition result is transmitted to the interface circuit 14. When the interface circuit 14 receives a recognition result from the character recognition circuit 13, it sends the recognition result to an external device (for example, an information processing device).When it receives a notification of an image signal abnormality from the character line detection circuit 12, it sends the image signal abnormality to an external device (for example, an information processing device). device). Because of this structure, recognition processing is possible if there is no code error in one character line, and recognition processing is possible if a line is missing due to a code error and the sub-scanning length of the character part is outside the allowable value. It is possible to determine that it cannot be performed normally and process it as a single image signal abnormality. Furthermore, even if the character information on the sheet cannot be read accurately due to an abnormality in the 5-fax facsimile, it can be processed based on the image signal abnormality.

As is clear from this result, it has become possible to significantly improve the accuracy of recognition results compared to conventional techniques.

(5) Description of Effects As explained above, according to the present invention, the original image signal disappears (line loss) due to a code error appearing in the character part of the sheet.
To determine whether normal recognition is possible by checking the number of scanning lines in the character part.

If there are many code errors in areas other than character lines, or if there are few code errors in character lines and do not affect recognition, recognition processing is possible; otherwise, it can be determined that the image signal is abnormal. Therefore, there is an advantage that more accurate recognition can be guaranteed than the conventional method in which image signal abnormality is determined simply based on the total number of code errors on one sheet or the total number of errors within a specified number of scanning lines. Also.

Even if character information on the input sheet is missing or extended due to an abnormality in the facsimile machine and cannot be read accurately, there is an advantage that recognition processing is not performed and the image signal is treated as an abnormality.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an embodiment of the present invention, FIG. 2 is a diagram showing an example of a sheet used in the present invention, and FIG. 3 is a diagram showing an example of an MH code decoding processing method. FIG. 4 is a diagram showing an example of a character line detection method. 1...Sheet, 2...Left character line designation mark, 3.
...Right character line designation mark, 4...Character entry frame, nl
-n5... Original image signal of each decoded scanning line, LT...
- The scanning line that detected the upper end of the left character line designation mark 2. RT...The scanning line that detected the top edge of the right character line designation mark 3, LB...The scanning line that detected the bottom edge of the left character line designation mark 2, RT...The scanning line that detected the bottom edge of the right character line designation mark 3 Scanning line whose lower end was detected, 10.14... Interface circuit, 11... Decoding circuit, 12... Character line detection circuit,
13...Character recognition circuit.

Claims (1)

[Scope of Claims] A character recognition method in which character information is written on a sheet on which a control mark indicating the location of the character information is written and inputted from a facsimile is transmitted and then recognized. A first step that decodes the encoded image signal of the sheet and, when a code error is detected, reproduces the original image signal in which the code error has occurred using the immediately previous original image signal.
means for detecting the position of the control mark in the sheet from the original image signal and measuring the length of the control mark; and using the length of the control mark measured by the second means to obtain the character information. and a fourth means for recognizing only the character information determined to be normally recognizable by the third means. Character recognition method using facsimile.