JPH07129706A - Method and device for character recognition processing and mark sheet used for them - Google Patents

Abstract

PURPOSE:To improve the recognition rate by performing correction to actual dimensions in the preprocessing of the character recognition processing and presenting high-quality facsimile data equivalent to a mark sheet original to a recognition processing part to reduce the probability of nonprescribed mark sheets. CONSTITUTION:After facsimile data is subjected to the preprocessing by a recognition preprocessing part 11, expansion/reduction rate detection start and end marks are detected and the interval between these marks is obtained in a subscanningdirection expansion/reduction rate detecting part 13. The value of this interval is compared with a prescribed value registered in a character recognition processing device 10 to obtain the expansion/reduction rate in the subscanning direction. The expansion/ reduction rate in the main scanning direction is obtained by the interval between right and left inclination detection marks of the mark sheet in a main scanning- direction expansion/reduction rate detecting part 14. It an actual dimension correction processing part 15, facsimile data is corrected to actual dimensions based on the expansion/reduction rates. Corrected facsimile data is subjected to recognition processing by a recognition processing part 12.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition processing system using a mark sheet or the like, and relates to an actual size of an input form such as a mark sheet or the size of data after reading, which is generated due to individual characteristics of a reading device such as a facsimile. Error is detected, the expansion / contraction ratio is accurately calculated, and the expansion / contraction ratio is corrected to the actual size. The present invention relates to a character recognition processing device with a correction function by expansion / contraction ratio detection that enables normal character recognition processing even outside the range, a character recognition processing device, and a mark sheet used therefor.

[0002]

2. Description of the Related Art Generally, a facsimile communication network provides a center end type facsimile communication service.

FIG. 7 shows the structure of a facsimile communication network. In FIG. 7, the facsimile machine 1 and the telephone 2 are connected to a computer 4 of an information processing center via a facsimile communication network 3. The facsimile communication network 3 includes therein a facsimile signal storage / conversion device (hereinafter abbreviated as STOC) 5 and a facsimile / data conversion connection device (hereinafter abbreviated as FDIC) for mutually converting a facsimile signal and a data signal. have.
Then, the facsimile machine 1 and the telephone 2 are connected to the computer 4 of the information center via the facsimile communication network composed of the STOC 5 and the FDIC 6.

Input from the facsimile device 1 is performed using a dedicated mark sheet, the FDIC 6 converts the facsimile signal into a code signal, and the converted signal is input to the computer 4.

8 and 9 show an example of a mark sheet used in this system. FIG. 8 shows a mark sheet A and FIG. 9 shows a mark sheet B. Here, the mark sheet A is a sheet on which the reference position mark or the like is printed in advance, and the mark sheet B is a sheet output from the facsimile communication network to the facsimile device.

The mark sheet has a reference position mark 7a, a start mark 7b, an end mark 7c, a mark sheet identification mark 7d, and a mark sheet which indicate the left end of the effective screen of the mark sheet, that is, the reference position in the main scanning direction when the mark sheet is input to the facsimile apparatus 1. Necessary for control at the time of reading such as inclination detection mark 7e for detecting the inclination of, a timing mark 7f corresponding to a number or a writing position of the mark, an information type designation mark for identifying three types of number, mark and pattern Marks (control marks) 7g, 7h, and 7i are printed with predetermined dimensions. In addition to these marks, a regular handwritten numeral designation mark 7j, a read skip mark row designation mark 7k, and a read skip mark 7l are similarly printed with predetermined dimensions.

Further, this mark sheet has a mark column 8
a, number field with guideline 8b, regular handwritten number field 8
c and a pattern entry field 8d are provided at predetermined positions.

In accordance with such control marks, the facsimile communication network 3 accurately reads the marks and numbers written on the mark sheet and converts them into code information. For example, when the mark sheet is inserted into the facsimile device 1 with an inclination, the FDIC 6 detects the inclination of the paper from the inclination detection mark of the received signal, electrically corrects the inclination, and then reads the number or mark. After that, it is controlled to transmit to the computer 4 of the information center.

[0009]

However, when the mark sheet A is transmitted from the facsimile apparatus 1 to the facsimile communication network 3, the characteristic of the facsimile causes expansion / contraction of the mark sheet A as compared with the actual size of the mark sheet A.
Therefore, when performing character recognition on the facsimile data received by the facsimile communication network 3, it may be considered that an unspecified mark sheet has been used and the recognition process cannot be executed depending on the degree of expansion / contraction.

FIG. 9 shows a conventional center delivery type mark sheet (mark sheet B) in a facsimile communication network. The mark sheet B is a mark sheet delivered from the information center connected to the facsimile communication network 3 to the facsimile device 1. The mark sheet B delivered from the information center to the facsimile device 1 expands and contracts due to the characteristics of the received facsimile.

Further, when the delivered mark sheet B is transmitted from the facsimile device 1 to the facsimile communication network 3,
Since expansion and contraction occur similarly to the above-described example, the actual expansion and contraction rate is twice that when the mark sheet A is used. Therefore, when character recognition is performed on the facsimile data received by the facsimile communication network 3, it may be considered that an unspecified mark sheet has been used and the recognition process cannot be executed depending on the degree of expansion and contraction.

Further, in the conventional mark sheet, when the expansion / contraction rate is to be grasped by the character recognition processing, the expansion / contraction can be detected by the reference value of the mark sheet in FIG. However, since each value has a narrow interval of several mm to several tens of mm, it is impossible to obtain a correctable expansion / contraction ratio.

Further, in the case of a mark sheet produced by copying a conventional mark sheet from an original with a copying machine or the like, expansion and contraction occur due to the characteristics of the copying machine. Therefore, when the copied mark sheet is used, the recognition device of the facsimile communication network handles the non-standard mark sheet, and the recognition process cannot be performed.

As described above, when the expansion / contraction rate detection function based on the characteristics of each facsimile is not provided, the following problems exist.

(I) When the mark sheet is out of regulation due to the expansion / contraction rate of the facsimile, the user cannot determine whether the mark sheet is erroneously designed or the mark sheet has a large error in printing. (Ii) Even if the expansion / contraction ratio of the facsimile is within the international standard, in the case of the center delivery type mark sheet, the expansion / contraction for the first time occurs when it is delivered to the facsimile, and further, when the delivered mark sheet is sent, Since the expansion and contraction occurs a second time, the expansion and contraction rate is doubled, and the expansion and contraction rate is outside the international standard. In the recognition processing, it is judged that the mark sheet is out of regulation and the recognition processing cannot be performed. (Iii) In a character recognition device using a conventional mark sheet, expansion / contraction can be detected, but since the specified value at each location for detection is small, a correctable expansion / contraction rate can be obtained. Can not. (IV) In the conventional mark sheet, the mark sheet made by copying from the original mark sheet by a copying machine etc. expands and contracts due to the characteristics of the copying machine, etc. Unable to process.

The present invention has been made in view of the above circumstances, and in the preprocessing in the character recognition processing in the character recognition processing apparatus, the actual size is corrected to obtain a high quality facsimile which is equivalent to the original mark sheet. An object of the present invention is to provide a character recognition processing method, a character recognition processing apparatus, and a mark sheet used for the same, which provide data to a recognition processing unit to reduce the probability of becoming a non-standard mark sheet and improve the recognition rate.

[0017]

That is, the present invention extracts mark information and numeral information from a facsimile image signal obtained by inputting from a facsimile apparatus a mark sheet having at least marks and numerals and converts the mark sheet into code information. In the character recognition processing method, a step of obtaining an interval between a pair of predetermined marks arranged in the sub-scanning direction of the mark sheet, and expansion / contraction of the facsimile image signal from an interval between the pair of predetermined marks on the facsimile image signal. It is characterized by comprising a step of obtaining a rate and a step of correcting the facsimile image signal to an actual size with the expansion / contraction rate.

Further, the present invention is a character recognition processing apparatus for extracting mark information and numeral information from a facsimile image signal obtained by inputting a mark sheet on which at least marks and numerals are written from a facsimile apparatus and converting the extracted mark information and numeral information into code information. In the above facsimile image signal,
A first means for detecting a pair of predetermined marks arranged at a predetermined interval in the sub-scanning direction of the mark sheet, and an interval between the pair of marks on the facsimile image signal detected by the first means. The second means for obtaining and the above first
Means for obtaining the expansion / contraction ratio of the facsimile image signal based on the interval detected by the means and the expansion ratio calculated by the second means, and the facsimile image signal based on the expansion / contraction ratio obtained by the third means. And a fourth means for correcting to the actual size.

Further, according to the present invention, in a mark sheet in which at least marks and numbers are written for character recognition processing, the expansion / contraction rate is set at a predetermined interval in the sub-scanning direction so that the expansion / contraction rate of the sheet can be detected. It is characterized by having a detection start mark and an expansion / contraction ratio detection end mark.

[0020]

According to the present invention, from the expansion / contraction ratio detection display on the mark sheet, the character recognition processing device grasps an accurate value capable of correcting the expansion / contraction ratio due to the characteristics at the time of reading such as fax, and further the character recognition processing. By correcting the mark sheet to the actual size with, the character recognition rate can be improved.

[0021]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing a configuration example of a character recognition processing device in a facsimile communication network according to the present invention.

In FIG. 1, a character recognition processing unit 10 includes a recognition preprocessing unit 11 which performs a preprocessing for recognizing characters and numbers from a facsimile image signal input from a facsimile device (not shown). A recognition processing unit 12 that performs recognition processing on the corrected facsimile data is provided. Further, in contrast to the conventional technique including the recognition preprocessing unit 11 and the recognition processing unit 12, in the present invention, the sub-scanning direction expansion / contraction rate detection unit 13, the main scanning direction expansion / contraction rate detection unit 14, and the actual size correction process. The part 15 is added to the configuration.

FIG. 2 and FIG. 3 show an example of the structure of the mark sheet with expansion / contraction ratio detection display used in this embodiment, and the correspondence between the constituent elements and the functional example of the mark sheet in FIG.

On the mark sheet, the left end of the effective screen of the mark sheet, that is, the reference position mark 7a indicating the reference position in the main scanning direction when input to the facsimile apparatus 1, the start mark 7b and the end mark 7c indicating the effective screen of the mark sheet. , The mark sheet identification mark 7d, the inclination detection mark 7e for detecting the inclination of the mark sheet, and the timing mark 7f corresponding to the number or the writing position of the mark,
It is printed with a predetermined size at a position as shown in FIG.

Further, on this mark sheet, marks (control marks) 7g, 7h, 7i necessary for reading control such as information type designation marks for identifying three kinds of numbers, marks and patterns are predetermined. It is printed with the dimensions. In addition to these marks, a regular handwritten numeral designation mark 7j, a read skip mark row designation mark 7k, and a read skip mark 7l are similarly printed with predetermined dimensions.

Further, between the timing mark 7f and the guideline-attached number line designation mark 7h, and between the mark line designation mark 7g and the guideline-attached number line designation mark 7h, the expansion / contraction ratio detection start position and end position are respectively indicated. The expansion / contraction ratio detection start mark 7m and the expansion / contraction ratio detection end mark 7n are printed with predetermined dimensions. The distance from the expansion / contraction rate detection start mark 7m to the expansion / contraction detection end mark 7n is an interval defined in the character recognition processing device.

Further, on this mark sheet, a mark column 8a, a guideline number column 8b, in which a mark, a number with a guideline, a regular handwritten number and a pattern are respectively entered,
A regular handwritten numeral column 8c and a pattern entry column 8d are provided at predetermined positions.

Next, referring to the flowchart of FIG.
A processing operation in the character recognition processing device in the facsimile communication network according to the embodiment will be described.

Facsimile data sent from a facsimile is first pre-processed by the recognition pre-processing unit 11.
Next, the expansion / contraction ratio detection unit 13 detects the expansion / contraction ratio detection start mark 7m (step S).
1). Subsequently, the expansion / contraction ratio detection end mark 7n is similarly detected (step S2). If both marks are detected in steps S1 and S2, the expansion / contraction ratio detection start mark 7m and expansion / contraction ratio detection start mark 7n are detected.
The spacing between is required.

The value obtained here is compared with the specified value registered in advance in the character recognition processing device 10 to obtain the expansion / contraction ratio in the sub-scanning direction. Subsequently, the main-scanning direction expansion / contraction ratio detection unit 14 obtains the expansion / contraction ratio in the main scanning direction from the interval between the left and right inclination detection marks 7e, which is one of the constituent elements of the mark sheet example shown in FIG. S
3).

Further, based on the expansion / contraction rate obtained by the actual size correction processing section 15, the actual size of the facsimile data sent from the facsimile is corrected (step S4). Then, the recognition processing unit 12 executes recognition processing on the corrected facsimile data (step S5), and this processing operation ends.

The character recognition processing device 10 uses the facsimile data of the mark sheet transmitted from the facsimile,
By comparing the interval from the expansion / contraction ratio detection start mark 7m to the expansion / contraction detection end mark 7n with a specified value, it is possible to accurately grasp the expansion / contraction ratio. Further, according to the expansion / contraction ratio, it is possible to perform correction to the actual size in the preprocessing of the recognition processing in the recognition preprocessing unit 11. Then, by passing the corrected data to the recognition processing unit 12, it is possible to reduce the recognition failure due to the non-standard mark sheet and improve the recognition rate.

The first actual size correction example in the sub-scanning direction will be described below.

Expansion and contraction in the sub-scanning direction in an optical reading device such as a facsimile is caused by a difference in width in which a mechanical portion such as a roller travels a sheet to be read from one reading to a next reading position. Characteristic arises. Therefore, when one mark sheet is read, expansion and contraction occur uniformly in the sub-scanning direction.

From this feature, the actual size correction in the sub-scanning direction is
This can be easily done by inserting or deleting the lines evenly with respect to the total number of lines of the mark sheet read by the expansion / contraction rate. At this time, when inserting, a copy of the previous line between the lines to be inserted is inserted.

FIG. 5 illustrates an example of actual size correction in the sub-scanning direction.

For example, it is assumed that the correction process is performed on the information shown in FIG. Now, it is assumed that the received facsimile image signal is reduced as shown in FIG. Then, the third shown in FIG.
As shown in FIG. 5C, the th line is inserted between the third line and the fourth line, and is corrected to the actual size as the 3'th line.

A second actual size correction example when a mark sheet copied by a copying machine or the like is used will be described.

The copied mark sheet has a property that expansion and contraction occur in the sub-scanning direction and the main scanning direction as compared with the original in view of the principle of the copying machine. The actual size correction in the sub-scanning direction is performed in the same manner as the above-described first actual size correction example.
The correction to the actual size in the main scanning direction is performed by inserting or deleting for each line, by finding the number of pixels to be inserted or reduced and the position based on the obtained expansion / contraction ratio. At this time, the determination of the white pixel / black pixel of the pixel to be inserted is the same as the previous pixel between the pixels to be inserted.

FIG. 6 shows another example of the structure of the mark sheet, showing a structure of a center delivery type mark sheet with expansion / contraction ratio detection display.

An expansion / contraction rate detection start mark 9a and an expansion / contraction detection end mark 9b are added and printed to the constituent elements of the conventional mark sheet. The interval from the expansion / contraction rate detection start mark 9a to the expansion / contraction detection end mark 9b is an interval defined in the character recognition processing device 10.

The character recognition processing device 10 determines, from the facsimile data of the mark sheet transmitted from the facsimile,
By comparing the interval from the expansion / contraction ratio detection start mark 9a to the expansion / contraction detection end mark 9b with a specified value, it is possible to accurately grasp the expansion / contraction ratio. Furthermore, it is possible to perform correction to the actual size in the preprocessing of the recognition processing based on the expansion / contraction ratio. Then, by passing the corrected data to the recognition processing unit, it is possible to reduce the recognition failure due to the non-standard mark sheet and improve the recognition rate.

[0044]

As described above, according to the present invention, the grasp of the expansion / contraction ratio due to individual characteristics of a facsimile or the like can be accurately grasped by the character recognition processing device, and therefore, the character recognition processing in the character recognition processing device can be performed. In the pre-processing in the above process, the actual size can be corrected and high quality facsimile data equivalent to the original mark sheet can be provided to the recognition processing unit. It is possible to improve.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration example of a character recognition device in a facsimile communication network in an embodiment of the present invention.

FIG. 2 is a diagram showing a configuration example of a mark sheet used in an embodiment of the present invention.

FIG. 3 is a diagram showing the correspondence between the constituent elements of the mark sheet of FIG. 2 and functional examples.

FIG. 4 is a flowchart illustrating a processing operation in a character recognition processing device in a facsimile communication network according to an embodiment of the present invention.

FIG. 5 is a diagram illustrating an example of actual size correction in the sub-scanning direction.

FIG. 6 is a diagram showing another configuration example of the mark sheet.

FIG. 7 is a diagram showing a configuration of a conventional facsimile communication network.

FIG. 8 is a diagram showing a configuration example of a conventional mark sheet used in a facsimile communication network.

FIG. 9 is a diagram showing another configuration example of a conventional mark sheet used in a facsimile communication network.

[Explanation of symbols]

7a ... reference position mark, 7b ... start mark, 7c ...
End mark, 7d ... Mark sheet identification mark, 7e ...
Tilt detection marks 7e, 7f ... Timing marks 7g ... Mark line designation mark, 7h ... Numerical line designation mark with guide line, 7i ... Pattern entry column designation mark, 7j ... Regular handwritten numeral designation mark, 7k ... Skip mark line designation mark, 7l ... skipping mark, 7m, 9a ... expansion / contraction ratio detection start mark, 7n, 9b ... expansion / contraction ratio detection end mark, 10 ... character recognition processing device, 11 ... recognition preprocessing unit, 1
2 ... Recognition processing unit, 13 ... Sub-scanning direction expansion / contraction rate detection unit, 14
... Main-scanning direction expansion / contraction rate detection unit, 15 ... Actual size correction processing unit.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display location H04N 1/393

Claims (3)

[Claims] 1. A character recognition processing method for extracting mark information and numeral information from a facsimile image signal obtained by inputting a mark sheet on which at least marks and numbers are written from a facsimile device and converting the extracted mark information and code information into code information. Determining the distance between a pair of predetermined marks arranged in the sub-scanning direction on the mark sheet; determining the expansion / contraction ratio of the facsimile image signal from the interval between the pair of predetermined marks on the facsimile image signal; And a step of correcting the facsimile image signal to an actual size. 2. A character recognition processing device for extracting mark information and numeral information from a facsimile image signal obtained by inputting a mark sheet on which at least marks and numbers are written from a facsimile device and converting the extracted mark information and code information into code information. First means for detecting a pair of predetermined marks arranged at a predetermined interval in the sub-scanning direction of the mark sheet on the facsimile image signal, and the facsimile image detected by the first means. Second means for obtaining the distance between the pair of marks on the signal, and third means for obtaining the expansion / contraction rate of the facsimile image signal based on the distance detected by the first means and the distance obtained by the second means. Character recognition means, and fourth means for correcting the facsimile image signal to an actual size based on the expansion / contraction ratio obtained by the third means. Apparatus. 3. A mark sheet in which at least marks and numbers are written to be subjected to character recognition processing, in order to detect the expansion / contraction rate of the sheet, an expansion / contraction rate detection start mark is provided at a predetermined interval in the sub-scanning direction. A mark sheet, which is provided with a stretch ratio detection end mark.