JPH03179583A - Character recognizing method - Google Patents

Abstract

PURPOSE:To detect a mixture recognition designation mark on the reception side and to decide a character area recognizing character type by using such a mark which can designate the plural recognizing character types by using an idle area at the right end of a start mark on an input sheet. CONSTITUTION:When the input sheet is inputted to a facsimile, it is inputted onto a picture signal memory like a figure. When the picture signal on the picture signal memory is investigated and the black picture signal of a reference position mark 2 is detected, the black picture signal of a start mark 3 at a specified position is detected from the upper end of the black picture signal of the mark 2 is detected. When the black picture signal of the mark 3 is detected, next, the black picture signals in mixture recognition designation mark columns 41-46 at prescribed positions are counted from the black picture signal of the mark 3 and when the prescribed number of black picture elements exist in respective areas, the presence of the description mark is decided. When detecting the black picture signals of one or both character line designation marks 6 and 7, according to the combination of the marks 6 and 7, the character type in a line is determined like a table and further, the combination for the presence/absence of the description mark in the mark columns 41-46 is decided like a table (F).

Description

[Detailed Description of the Invention] (Industrial Application Field) This invention relates to a character recognition method that performs character recognition by determining two character types in a character area from various control marks on an input form input by facsimile, etc. It is.

[Conventional technology]

Conventionally, in order to accommodate a large number of users and services, recognition devices installed in remote locations do not have format information on the recognition device side, but instead provide various control marks on input paper 1 as shown in FIG. Detecting these control marks made it possible to process various formats. In particular, it is better to limit the character types to be recognized as much as possible to increase the correct reading rate, and character line designation marks 6, 7, and 8 can be arranged in the main scanning direction to identify numeric lines, alphabetic character lines, and kana lines (character type determination was a combination of character line designation marks 6 and 7).

[Problem to be solved by the invention]

However, depending on the service, there may be numeric lines or alphabetic lines.

In addition to kana lines, lines with various combinations are required, such as lines with numbers, letters, and letters, and lines with numbers, letters, and kana, even if the orthodoxy rate is slightly lower. If a character line designation mark that enables recognition of lines with multiple character types is added using the conventional method, the area for writing the character line designation mark will increase.
There is a problem in that the area in which text information is written by the user must be reduced. Further, there is a problem in that the recognition device cannot recognize the conventional input paper 1 by simply adding a function for determining the added character line designation mark.

This invention was made to eliminate these problems, and aims to enable a mixed recognition service without reducing the character information area written by the user, and also to enable the use of conventional input forms. .

[Means for Solving the Problems] The character recognition method according to the present invention includes a start mark at the top end of the entire character information, an end mark at the bottom end, and a character line designation indicating the position and width of the character information in the sub-scanning direction. The mark is entered using an input sheet with multiple mixed recognition designation marks written in a color that can be detected by the sensor at a position adjacent to the start mark in the main scanning direction. The start mark, character line designation mark, and enmark are detected from the image signal detected, and the presence or absence of multiple mixed recognition designation marks that exist in the main scanning direction of the start mark is detected, and mixed recognition is performed from the detected multiple mixed recognition designation marks. The character type and mixed recognition target area are determined, the character pattern is extracted from the image information of the input paper based on the position of the detected character line designation mark, and the extracted character pattern is based on the detected character type. It recognizes the characters and converts them into character codes.

[Effect]

In this invention, when specifying the mixed recognition area and mixed recognition character types on the input paper, the mixed recognition area and mixed recognition character types are specified using the empty area near the start mark used conventionally. , which allows the recognition device to identify mixed recognition areas and mixed recognition character type instructions, and the character information writing area is larger than the conventional method of increasing control marks in the main scanning direction to correspond to recognized character types. , there is a degree of freedom in specifying mixed recognition areas, and conventional input forms can also be processed.

(Example) FIG. 1 is a diagram showing an example of an input sheet used in the present invention, where 1 is an input sheet, 2 is a reference position mark indicating the left edge of this input sheet 1, and 3 is an upper end of the input sheet 1. or a start mark indicating the upper end of the entire character information; 4 is a mixed recognition designation mark field; 5 is an end mark indicating the lower end of the input paper 1 or the lower end of the entire character information; 6, 7.8 are the sub-scanning direction of the character line Character line designation mark indicating the existing position 1 width, 9
is a character entry field in which character information is entered.

Note that the reference position mark 2. Start mark 3 End mark 52 Character line designation mark 6.7.8 is called a control mark and is written in a color that can be detected by the facsimile sensor, and mixed recognition designation mark field 4 and character entry frame 9 are written using the facsimile sharp sensor. Described with an imperceptible dropout color. However, the sizes of these various control marks are defined in advance, and the positions in the main scanning direction from the reference position mark 2 are also defined. In addition, the size and arrangement of the mixed recognition designation mark field 4 and the character entry frame 9 are also specified.

Figure 2 is a diagram showing the detection and character type determination processing of various control marks, Figure 2 (a) is a diagram showing the relationship and detection method of various control marks, and Figure 2 (b) is a diagram showing the character line specification. mark 6
．． FIG. 2(C) is a diagram showing a method for determining character type from 41 to 7, and FIG. 46
is the mixed recognition designation mark field, 41.42 is mixed designation for numeric rows, 43.44 is mixed designation for alphabetic rows, 4
5.46 performs mixed specification for kana lines.

Next, a method for detecting various control marks and determining character types will be explained using FIG. First, when input paper 1 is input into the facsimile machine, the image signal memory shown in FIG. 2(a)
is entered as shown in .

Then, the image signal on the image signal memory is checked, and the black image signal of the reference position mark 2 is detected. When the black image signal of the reference position mark 2 is detected, the black image signal of the start mark 3 located at a specific position from the upper end of the black image signal of the reference position mark 2 is detected. When the black pixel signal of start mark 3 is detected, the black pixel signals within the mixed recognition designation marks m41 to m46 located at predetermined positions are counted from the black pixel signal of the start mark, and the black pixel of each area is If the number of marks exists, it is determined that there is an entry mark. When the presence or absence of entry marks in the mixed recognition designation mark fields 41 to 46 is determined, the character line designation mark 6.7 located at a specific position is then determined from the black image signal of the reference position mark 2.
．． 8 and end mark 5 are detected.

When either or both of the character line designation marks 6.7 and black image signals are detected, the in-line character type is determined by the combination of the character line designation marks 6.7, as shown in FIG. 2(b). When the black image signal of any of the character line designation marks 6.7 is detected, the character at a specific position considering the allowable inclination is determined from the black image signals of the reference position mark 2 and the character line designation mark 6.7. A black image signal of the line designation mark 8 is detected. When a character line is detected, the previously determined combinations of entry mark presence/absence in the mixed recognition designation mark columns 41 to 46 are determined as shown in FIG. 2(C).

In this embodiment, the mixed recognition designation mark field 43°45.46
If it is determined that there is an entry mark in input form 1, the 1st and 2nd lines of input form 1 will recognize single numbers, the 3rd line will recognize a mixture of alphabetic characters and kana, and the 4th line will recognize mixed character types of numbers and alphabetic characters. becomes.

Finally, when the black image signal of the end mark 5 at a specific position is detected from the black image signal of the reference position mark 2, the detection of various control marks from the input paper 1 ends. However, when detecting various control marks, if a black image signal of a specified size is detected at a specified position, it is determined that a control mark is present.

FIG. 3 is a diagram showing the configuration of an example of an apparatus for carrying out the present invention, in which 100 is a fuse for inputting the input paper 1, 101 is for finishing the facsimile transmission procedure and extracting the image signal from the facsimile signal. A facsimile interface unit, 102 is an image signal memory unit that stores image signals, 103 is a control mark detection/character type determination unit that detects various control marks and determines character types, and 104 extracts character stamps from image signals. 105 is a character recognition unit that recognizes a character slam and converts it into a code; 106 is an information processing device interface unit that transfers the code signal of the recognition result to the information IA management device; 107 is a processing device that processes the recognition result. This is an information processing device.

Next, the operation shown in FIG. 3 will be explained. First, input paper 1 is input into facsimile 100. Facsimile 10
0 scans the input paper 1 and transmits a facsimile signal to the facsimile interface 111101 via the line. The facsimile sharp ink face section 101 receives signals from the facsimile, and extracts an image signal from the facsimile signal and stores it in an image signal memory section 102. When the image signal has been stored, the control mark detection/character type determination unit 103 is notified of the completion of reception. Upon completion of reception, the control mark detection/character type determination section 103 reads out the image signal from the image signal memory section 102, and detects the reference position mark 2 and start mark 3 using the method shown in FIG. start mark 3
When detected, the character slam extraction unit 1041 character recognition unit 1
05. An operation start instruction is given to the information processing device interface unit 106. Then, it is determined whether entries have been made in the mixed recognition designation mark columns 41 to 46 using the method described in FIG. Next, the positions of the reference position mark 2, the character line designation mark 6, 7, and 8 of the first line are detected by the method shown in FIG. At the same time, it instructs to extract one line of character stamps. Also,
The recognition chamber ti of the character line is determined by the method shown in FIGS. 2(b) and 2(c). The character slam extraction unit 104 extracts a character entry frame 9 located at a predetermined position from the position of the reference position mark 2 and the character line designation mark 6.7.8 on the image signal.
The image signal of the character bang is extracted and transferred to the character recognition unit 105.

When character stamps for one line have been extracted and transferred, the control mark detection/character type determination unit 103 is notified of completion of character stamp extraction. The control mark detection/character type determination unit 103 notifies the character recognition unit 105 of the recognized character type for one line and instructs the character recognition unit 105 to start recognition. The character recognition unit 105, which has received the recognition character type and the instruction to start recognition, recognizes the received character slam image signal as the designated character type, converts it into a code signal, and transfers it to the information processing device interface unit 106. When the recognition and transfer of one line is completed, the character recognition unit 105 notifies the control mark detection/character type determination unit 103 of the completion of recognition. Control mark detection/control mark receiving notification of one line of character stamp extraction completion and recognition completion
The character type determination unit 103 issues instructions to the character slam extraction unit 104 and the character recognition unit 105 in the same manner as in the first line to continue processing.

Then, the end mark 5 is detected by the control mark detection/character type determination unit 103, and the character recognition unit 105 indicates that the recognition of the last line has been completed.
When notified from the information processing device interface unit 10
6 to transfer the recognition results. The information processing device interface unit 106 inputs the code signal (character recognition result) of the input paper 1 received from the character recognition unit 105 to the information processing device 1.
07, and when the transfer is completed, the control mark detection/character type determination unit 103 is notified of the transfer completion.

During the above series of processing, the control mark detection/character type determination unit 10
If the control mark cannot be detected normally in step 3, the information processing device interface unit 106 is notified of abnormal termination after the character recognition unit 105 has completed the requested recognition. The information processing device interface unit 106 that has received the abnormal termination transmits the processing result and information to the effect that the abnormal termination has occurred to the information processing device 107, and when the transfer is completed, notifies the control mark detection/character type determination unit 103 of the completion of the transfer. .

In the above embodiment, a mixed recognition designation mark field is provided and entered by the user, but it may be printed in advance. In addition, the mixed character type specification method is numeric line, alphabetic line,
Because it was specified separately from the kana line, six mixed character type specification and constant mark fields were required, but it is also possible to have three mixed character type specification mark fields for numbers, alphabets, and kana lines, or if necessary. All you have to do is decide on the combination. Furthermore, the mixed recognition designation mark field can be used to allow mixed recognition designation of a specific area of multiple character lines including two-column line information.

〔Effect of the invention〕

As explained above, this invention uses a mixed recognition designation mark that allows multiple recognition character types to be specified using the empty area at the right end of the start mark on the input paper, and the recognition device side that is the receiving side Since the recognition designation mark can be detected and the character type recognized in the character area can be determined, even when designing an input form where multiple character types exist in a character line, it is possible to
Compared to the conventional method, this method has the advantage that the text information writing area written by the user can be set larger. In addition, even in the case of currently used input forms, there is the advantage that mixed recognition can be specified for character types other than the currently specified recognition character types without changing the format (simply by adding character line designation marks to empty areas). . Furthermore, even if functions for detecting mixed recognition designation marks and determining character types are added to the recognition device, there is an advantage that services using conventional input forms can also be performed.

[Brief explanation of drawings]

Figure 1 is a diagram showing an example of an input form used in this invention, Figure 2 is a diagram showing detection and determination processing of various control marks, and Figure 2 (a) is a diagram showing the relationship and detection of each control mark. Figure 2 (b) is a diagram showing how to determine the character type of a character line from the character line designation mark, and Figure 2 (C) is a diagram showing how to determine the character type of a character line from the character line designation mark and the mixed recognition designation mark. FIG. 3 is a block diagram showing the configuration of an embodiment of the present invention, and FIG. 4 shows an example of an input form used in a conventional character recognition device. It is a diagram. In the figure, 1 is input paper, 2 is reference position mark, 3 is start mark, 4 is mixed recognition designation mark field, 5 is end mark, 6.7.8 is character line designation mark, 9 is character entry frame ,
41 to 46 are mixed recognition designation mark fields, 100 is a facsimile, 1o1 is a facsimile interface section, 102
103 is a control mark detection/character type determination unit; 104 is a character slam extraction unit; 105 is a character recognition unit;
10th is an information processing device interface unit, and 107 is an information processing device. Figure Figure (a) 1-46 Mizai Textile Pilgrimage T Mark Kyouran Figure 4

Claims (1)

[Claims] In a character recognition method in which character information on an input form whose image is input from a transmitting side is recognized on a receiving side, a start mark is placed at the top end of the entire character information, an end mark is placed at the bottom end, and the presence of the character information in the sub-scanning direction. This input form from the sending side uses an input paper on which a character line designation mark indicating the position and width is written in a color that can be detected by a sensor, and a plurality of mixed recognition designation marks are written in a position in the main scanning direction adjacent to the start mark. Detects the start mark, the character line designation mark, and the encode mark, and determines the presence or absence of multiple mixed recognition designation marks that exist in the main scanning direction of the start mark from the image signal obtained on the receiving side. A mixed recognition character type and a mixed recognition target area are determined from the plurality of detected mixed recognition designation marks, and a character pattern is determined from the image information on the input paper based on the position of the detected character line designation mark. A character recognition method comprising: extracting a character pattern, recognizing the extracted character pattern based on the detected character type, and converting it into a character code.