JPH07296108A - Method and device for recognizing character - Google Patents

Abstract

PURPOSE:To provide the character recognizing method with which described data can be recognized even without using any dedicated OCR paper and a format change accompanying the change of a described data item can be easily performed at low cost. CONSTITUTION:This device is composed of an image data input part 21, picture element deciding part 22 for deciding white and black picture elements for each line of inputted image data, black picture element counter 23 for counting the number of black picture elements in the line in a prescribed direction, black picture element comparing part 24 for comparing the amount of prescribed continued black picture elements with the value of the black picture element counter 23, frame recognizing part 26 for deciding whether the scanned lines express a frame or not according to the result of the black picture element comparing part 24, and character recognizing part 27 for recognizing the characters described inside that frame.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition method and a character recognition apparatus, and more particularly to a character recognition method and a character recognition apparatus for detecting a frame described on a sheet and recognizing a character within the frame.

Normally, the entry fields on a predetermined OCR entry sheet are designed according to the business pattern and ordered from a printing company. However, since this OCR entry sheet is expensive, it is a normal printing device such as a word processor. There is a demand for a method of using plain paper as a data registration paper by using.

[0003]

2. Description of the Related Art FIG. 10 is a diagram for explaining a conventional character recognition process. An example is shown in which the image data shown in the figure is written on an OCR sheet and the image data input by facsimile communication is recognized. In advance, "A, B, C, D"
In the field marked with, scan the location set at n from the left edge of the paper and m from the top edge,
In the field in which "1" is entered, the OCR definition information is set so that scanning is performed from the position r from the left end of the sheet and o from the upper end. In this way, when the image data of the pattern shown in the figure is input, the preset position is scanned and the character described at the predetermined position is recognized. Although a frame is shown in the figure, the frame is not actually output on the image data transmitted by facsimile because the frame is dropped out.

Further, if the paper is tilted, the position of the frame cannot be accurately recognized, so a method of positioning and scanning using a correction mark or the like is used.

The case of scanning characters using these methods will be described. FIG. 11 is a diagram for explaining a conventional OCR recognition and correction procedure. First, the vertical correction of the image data input by the facsimile mark 200 and the facsimile correction mark 300 is performed. Next, the character recognition position is determined by the OCR definition information. When character-recognizing the field a, first, the definition information (start position, length) of a is acquired from the OCR definition information, and the character recognition position is determined. In this case, it is positioned at the coordinate position (2, 3). Next, the character recognition processing is mechanically performed in order for the specified length of the character frame from the character recognition start position.

The above processing is performed for all fields in the OCR definition information.

[0007]

However, in the above-described conventional character recognition method, the field to be scanned is defined in advance by the OCR definition information, and the recognition process is performed based on the information, so that the contents of the work may be changed. When the OCR entry data item is changed, the OCR definition information must be changed. Also, the OCR entry form must be ordered from the printing company as the entry data items are changed. This OCR entry sheet also has a problem that the cost is high because the printing of entry fields is special printing.

The present invention has been made in view of the above points, solves the problems of the prior art, and can recognize the entered data without using a dedicated OCR sheet, the format accompanying the change of the entered data items. An object of the present invention is to provide a character recognition device and a character recognition method that can be easily changed at low cost.

A further object of the present invention is to provide a character recognition device and a character recognition method capable of detecting the specific frame and using the characters described in the frame for arbitrary processing.

[0010]

The present invention, when scanning and recognizing a character written in a specific frame of input image data, does not define a position and a length to be read in advance, When the image data is scanned and the black pixel arrangement is more than a predetermined value and the black pixel arrangement constitutes a frame, it is recognized as a specific frame, and the characters entered in the specific frame are recognized. Recognize.

The character recognition method of the present invention is a method of inputting characters or characters entered in a specific frame of image data,
In the symbol recognition method, operate the image data,
A portion in which a predetermined arrangement of pixels constitutes a specific frame is recognized, and the recognized pixel in the specific frame is recognized as a character or a symbol written in the specific frame.

Further, the character recognition method of the present invention operates the image data in the image data processing method in which arbitrary processing is performed based on the characters or symbols written in a specific frame of the input image data. Recognize the part where the predetermined pixel arrangement constitutes a specific frame, recognize the pixels in the recognized specific frame as the characters written in the specific frame, and the appearance order of the recognized specific frame Then, arbitrary processing is performed based on the recognized characters in the frame.

Further, the character recognition device of the present invention is a character recognition device for recognizing a character entered in a specific frame of input image data, by scanning the image data to specify a predetermined pixel arrangement. The frame recognizing means for recognizing the portion forming the frame and the character recognizing means for recognizing the pixels in the specific frame recognized by the frame recognizing means as the characters written in the specific frame.

Further, the character recognition apparatus of the present invention is an image data processing apparatus which performs an arbitrary process based on a character entered in a specific frame of input image data, and scans the image data to determine a predetermined value. Frame recognition means for recognizing a portion in which the arrangement of pixels constitutes the specific frame, and character recognition for recognizing pixels in the specific frame recognized by the frame recognition means as characters written in the specific frame And a processing unit that performs an arbitrary process based on the appearance order of the frame recognized by the frame recognition unit and the characters in the frame recognized by the character recognition unit.

[0015]

According to the present invention, even if the scanning position is not set in advance like the OCR definition information, a frame for writing characters is provided by a word processor or the like, and the user writes data in the writing frame. Then, when scanning with a scanner or the like, the horizontal axis and the vertical axis where a predetermined number or more of black pixels are continuous are determined, and it is recognized that the part is a character frame in which a character is written, and within that character frame The data can be recognized by reading the written characters. Therefore, even if the OCR dedicated paper printed with special ink is not used, the character frame for entering the data is printed by the user or the word processor etc. at the time of facsimile transmission, and the data is written in the frame. Fill in and send by fax. On the receiving side, the written data can be read by scanning the character frame field of this image data.

Further, by defining the characters, symbols, etc. described in a specific frame, it becomes possible to use them for various processes such as application programs.

[0017]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIG. 1 shows the system configuration of an embodiment of the present invention. The system shown in the figure recognizes the character frame of the image data received from the facsimile machine 10 by the user writing the data on the entry sheet created on the plain paper by the word processor, transmitting the data by the facsimile machine 10 on the transmitting side. The recognition device 20 recognizes the characters entered in the frame. In addition, hereinafter, the characters written in the frame are displayed, but they are simply marks.
Is also included.

The recognition device 20 includes an image data input unit 21 such as a scanner for scanning image data, a pixel determination unit 22 for determining white pixels and black pixels for each line of the input image data, and a line in a predetermined direction. According to the results of the black pixel counter 23 that counts the number of black pixels, the black pixel comparison unit 24 that compares the value of the black pixel counter 23 with the predetermined continuous amount of black pixels stored in the work area 25, and the result of the black pixel comparison unit 24. A frame recognition unit 26 that determines whether or not the scanned line represents a frame, and a character recognition that recognizes a character entered in the frame when the frame recognition unit 26 determines that the line represents a frame. It is composed of a section 27.

FIG. 2 is a diagram for explaining character recognition according to an embodiment of the present invention. Unlike the conventional method, the line mark is unnecessary. First, the vertical correction of the image data input by the facsimile mark 200 and the facsimile correction mark 300 is performed. The character recognition field (black frame) is detected from the image data by the continuous amount of black pixels. The presence or absence of a character frame (black frame) is searched in order from the X direction to the Y direction.
Performs character recognition processing. The above processing is performed until there is no character frame on the image information.

FIG. 3 is a flow chart showing the outline of the character recognition field detecting operation of the embodiment of the present invention, and FIG. 4 shows the scanning direction of the embodiment of the present invention.

Step 1) The image of the first input image data is expanded.

Step 2) When the scanning of the field data in the image data is completed, the processing is completed.

Step 3) When a black pixel is detected in the field data of the image data, first, the X axis is scanned in the b direction, and the continuous amount of black pixels is counted every time the black pixel is detected.

Step 4) The number of consecutive black pixels detected when scanning in the X-axis direction b being counted is compared with a predetermined value.

Step 5) As a result of the comparison, if it is equal to or more than a predetermined number, it is determined to be continuous in the X-axis direction, and then the line in the Y-axis direction c continuous to the X-axis is scanned. Similar to the above, every time a black pixel is detected in the Y-axis direction c, the continuous amount of black pixels in the Y-axis direction c is counted.

Step 6) Here, the number of black pixels detected when scanning in the counted Y-axis direction b is compared with a predetermined value.

Step 7) As a result of the comparison, if the number is equal to or more than a predetermined number, it is determined that they are continuous in the Y-axis direction c, and the Y
The X axis, which is continuous with the axis, is scanned in the direction d. Similar to the above,
The scanning is performed in the X-axis direction d, and the continuous amount of black pixels in the X-axis direction d is counted every time a black pixel is detected.

Step 8) The number of black pixels detected when scanning in the counted X-axis direction d is compared with a predetermined value.

Step 9) As a result of the comparison, if the number of black pixels is equal to or more than a predetermined number, it is determined that they are continuous in the X-axis direction d,
The Y-axis e which is continuous with the X-axis is scanned. Similar to the above,
Every time a black pixel is detected in the Y-axis direction e, the continuous amount of black pixels in the Y-axis direction e is counted.

Step 10) In the above process, when the X-axis in the two directions and the Y-axis in the two directions are all continuous, the reading field is determined and the character written in the field is read.

Step 11) The recognition processing of the read character is performed.

Next, an example of scanning the X-axis and the Y-axis and counting the number of pixels will be described. FIG. 5 shows an example of counting black pixels according to an embodiment of the present invention. In the figure, the preset black pixel continuous amount is 3 bits, "1" indicates a black pixel, and "0" indicates a white pixel. Therefore, if at least 3 bits of black pixels are not continuous, it is determined not to be a character frame.

The count is incremented while the black pixels are continuous, and is zero-cleared in the case of a discontinuous state.

At this time, in FIG. 9A, the bit arrangement in the direction b of the X axis is "11110111", and the black pixel continuous amount is 4 bits or 3 bits, which is equal to or equal to the predetermined continuous amount. Since it is above, it is determined that they are continuous. The bit arrangement in the direction c of the Y-axis is “1110111101”, and the black pixel continuous amount is 3 bits or 4 bits, which is equal to or larger than the predetermined continuous amount, so that it is determined to be continuous. . The bit arrangement in the direction d of the X axis is “11101111”, and the continuous black pixel amount is 3 bits or 4 bits, which is equal to or larger than the predetermined continuous amount, and thus is determined to be continuous. . The bit array in the direction e of the Y-axis is “111110111”, and the black pixel continuous amount is 5 bits and 3 bits, so it is determined that the black pixels are continuous. Therefore, the field shown in FIG. 9A is recognized as a character frame.

In FIG. 7B, the bit arrangement in the direction b of the X axis is "01101111", and the continuous black pixel amount is 2 bits or 4 bits, which is larger than the predetermined continuous amount, so that the continuous black pixels are continuously generated. It is determined that
The bit arrangement in the direction c of the Y axis is “1100111011”, the black pixel continuous amount is 2 bits, 3 bits, and 2 bits, and there is a continuous value that is the same value as the predetermined continuous amount, so that it is continuous. To be judged. The bit arrangement in the direction d of the X-axis is “1110011”, the black pixel continuous amount is 3 bits and 2 bits, and since there is a continuous amount having the same value as the predetermined continuous amount, it is determined to be continuous. The bit arrangement in the direction e of the Y-axis is “1011100110”, and since the black pixel continuous amount is 3 bits and 2 bits, there is a continuous amount of the same value as the predetermined continuous amount, so it is determined to be continuous. . Therefore, in the field shown in FIG. 6B, the number of consecutive bits is smaller than that in the above-mentioned (A), but since there is a number of consecutive bits equal to or more than a predetermined continuous amount,
Recognized as a character frame.

In FIG. 7C, the bit array in the direction b of the X axis is "10110010", and since the continuous amount of black pixels is only 2 bits, the black pixels in the direction b of the X axis at this point. The line is determined not to be continuous and is not recognized as a character frame.

As another embodiment, there is also a method of determining whether or not a pattern having a predetermined continuous amount is repeated when a character frame is printed with a dotted line. FIG. 6 shows a case where the character frame of another embodiment of the present invention is a dotted line. For example, the continuous amount of predetermined black pixels and the continuous amount of white pixels are set to 2 bits each. That is, in the case of "1100" or "0011", it is possible to recognize as a character frame surrounded by a dotted line even if only black pixels are not continuous. For example, the count in this case is "110.
When "0" or "0011" appears, the count is performed, and when at least two patterns appear, the character frame is surrounded by a dotted line. In the example of FIG. The bit array is “1100001101”, and there are “1100” and “0011”, so it is determined to be a dotted line.Next, in the scanning in the c direction of the Y axis, the bit array is “1100011001” and “1100”. Since there is "1100", X axis b
Judge as a dotted line continuous in the direction. In the scanning in the X-axis d direction, the bit array is “11000011001” and there are “1100” and “0011”, so that the Y-axis c
Judge as a dotted line continuous in the direction. In scanning in the Y-axis e direction, the bit array is "1100100111", and there are "1100" and "0011", so the X-axis d
Judge as a dotted line continuous in the direction.

Further, in the above description, the continuous amount of black pixels may simply be designated as 2 bits or more. Further, several kinds of bit continuous patterns may be set. In the example of FIG. 6, since the part where bit 1 stands is a black pixel,
The above "1100" and "0011" are divided in 4-bit units, but the character frame may be determined by dividing the n-bit units and the number of times the pattern appears. For example, when divided into 3 bits, "111", "111", "110",
When "100", "011", and "001" appear, it is possible to determine that the character frame is surrounded by a dotted line.

In the above embodiment, when counting the number of consecutive black pixels, each side of the frame was scanned clockwise, but which side is the starting point and in which direction the scanning is optional. .

Further, a parity check may be used to detect the continuous or discontinuous state of black pixels.

Further, FIG. 8 is a diagram for explaining the second embodiment of the present invention.

In this embodiment, in the above embodiment, a specific frame is detected, the frame is defined, and the data in the frame is treated as code conversion data. As a result, it is possible to use the character data within the recognized character frame with the unique name defined by the application program. The figure shows an ordering slip and a state in which the data entered in each frame of the ordering slip is edited.

FIG. 9 is a flow chart showing the operation of the second embodiment of the present invention until the editing after the recognition. The operation of the following flowchart will be described based on the example of FIG.

The three retrieved frames are associated with the product name code, and the three retrieved frames are associated with the quantity (step 101).

Next, the recognition result is input (step 102), it is judged whether the recognition item is an alphanumeric item or a numeric item (step 103), and the product name is an alphanumeric item. Search the table (step 10
4) It is determined whether a matching product name code exists, and if it exists (step 105), the product name is obtained from the product name table (step 106).

If the recognition item is a numeric item (step 103), a numeric check is performed (step 107). If there is no error, the numeric value of the numeric item is acquired (step 108) and the product name is entered. The data for one record is edited (step 109). The above process is performed up to the last line, and when the process of the last line is completed, the total number of items, the total of the ordered quantity, etc. are calculated, and the data editing is completed (step 111).

The present invention is not limited to the above-described embodiments, and various applications and modifications are possible within the scope of the claims.

[0049]

As described above, according to the present invention, the continuous amount of black pixels is counted, and when a line having a continuous amount of a predetermined value or more is scanned in four directions, it is recognized as a character frame, It is possible to perform the recognition process of the characters in the character frame. Therefore, even if a dedicated OCR sheet is not used, an ordinary word processor or a user writes characters in a frame created on plain paper to perform facsimile communication,
Since the recognition processing can be performed, the cost is reduced. In addition, like a dedicated OCR sheet, every time the work content is changed, a new OCR sheet is redesigned and data can be input in an arbitrary format by a word processor or user without ordering printing from a vendor. It will be possible.

Further, it is possible to use the characters, symbols and the like described in the recognized specific frame for various jobs by defining a unique name for various processes such as an application.

[Brief description of drawings]

FIG. 1 is a system configuration diagram of the present invention.

FIG. 2 is a diagram for explaining character recognition according to an embodiment of the present invention.

FIG. 3 is a flowchart of a character recognition field detecting operation according to an embodiment of the present invention.

FIG. 4 is a diagram showing a scanning direction according to an embodiment of the present invention.

FIG. 5 is a diagram showing an example of counting black pixels according to an embodiment of the present invention.

FIG. 6 is a diagram showing a case where a character frame according to another embodiment of the present invention is a dotted line.

FIG. 7 is a diagram showing an example of uniquely specifying a data item in a character frame to an application according to an embodiment of the present invention.

FIG. 8 is a diagram for explaining the second embodiment of the present invention.

FIG. 9 is a flowchart showing an operation until recognition and editing in the second embodiment of the present invention.

FIG. 10 is a diagram for explaining a conventional character recognition process.

FIG. 11 is a diagram for explaining a conventional OCR recognition and correction procedure.

[Explanation of symbols]

 10 Facsimile device 20 Recognition device 21 Image data input unit 22 Pixel determination unit 23 Black pixel counter 24 Black pixel comparison unit 25 Work area (predetermined count) 26 Frame recognition unit 27 Character recognition unit 100 line mark 200 Facsimile mark 300 Facsimile correction mark

Claims (4)

[Claims] 1. A method of recognizing a character or a symbol written in a specific frame of input image data, wherein the image data is manipulated so that a predetermined pixel arrangement constitutes the specific frame. A character recognizing method, characterized in that a part in which a particular part is recognized is recognized, and a pixel in the recognized specific frame is recognized as a character or a symbol written in the specific frame. 2. An image data processing method for performing an arbitrary process based on a character or a symbol entered in a specific frame of input image data, wherein the image data is operated to arrange a predetermined pixel array. Recognizes a portion forming the specific frame, recognizes a pixel in the recognized specific frame as a character written in the specific frame, and recognizes the appearance order of the recognized specific frame. A character recognition method characterized by performing arbitrary processing based on characters in a frame. 3. A character recognition device for recognizing a character entered in a specific frame of input image data, wherein the image data is scanned and a predetermined pixel array constitutes the specific frame. Character recognition characterized by having a frame recognizing means for recognizing an existing portion and a character recognizing means for recognizing pixels in a specific frame recognized by the frame recognizing means as characters written in the specific frame apparatus. 4. An image data processing apparatus for performing an arbitrary process based on a character entered in a specific frame of input image data, wherein the image data is scanned and a predetermined arrangement of pixels has the specific pixel arrangement. A frame recognition means for recognizing a part forming a frame, a character recognition means for recognizing pixels in a specific frame recognized by the frame recognition means as characters written in the specific frame, and the frame recognition A character recognition apparatus comprising: a processing unit that performs an arbitrary process based on the appearance order of the frame recognized by the means and the characters in the frame recognized by the character recognition unit.