JPH04351064A - Facsimile business form processing method - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention relates to a facsimile document processing method for recognizing characters, tables, figures, etc. on a document based on image data on the document transmitted from a facsimile machine, and particularly relates to a facsimile document processing method that recognizes characters, tables, figures, etc. The present invention relates to a facsimile form processing method suitable for enabling highly accurate recognition processing even for forms that are not provided with edge marks for accuracy correction.

0002

[Prior Art] In a facsimile OCR system that OCR recognizes characters (alphanumeric characters, kana, symbols, kanji, etc.) on a form based on the image data of the form sent from a facsimile, the conventional technology is as shown in FIG. A form 1 with pre-printed edge marks 1a as shown is used. With the form 1 on which such edge marks 1a are pre-printed, the OC
Before character cutting processing, which is preprocessing for R recognition, the number of white dots between two edge marks 1a in the horizontal and vertical directions is calculated, and the scanning accuracy of the facsimile used for image input is calculated and corrected. This increases character extraction accuracy and improves recognition rate.

[0003]In connection with the above-mentioned prior art, there is an invention disclosed in Japanese Patent Publication No. 1-35381.

0004

[Problems to be Solved by the Invention] In the above-mentioned prior art, only the edge-marked form 1 shown in FIG. 2 can be used;
There was a problem in that a normal OCR form 2 without edge marks as shown in FIG. 3 could not be used.

Furthermore, in the case of a form recognition system that recognizes tables (cross lines, frames), characters, etc. on a form based on the image data, the format of the already printed form is recognized. Therefore, in the case of the form 2 without edge marks printed, it is impossible to correct the scanning accuracy of the facsimile machine, so there is a problem that the recognized pitch between the lines and the frame size are incorrect. In this case, we had no choice but to limit and select the types of facsimile machines to be used.

As described above, in the conventional facsimile form processing system, in order to correct the scanning accuracy of the facsimile, edge marks for correcting the scanning accuracy are provided on the form and the number of dots between the edge marks is measured. The main and sub-scanning accuracy of the facsimile was measured and corrected. Therefore, as a matter of course, there was a problem that the only usable forms were forms with edge marks.

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems, and to provide a facsimile form processing method that makes it possible to use forms without edge marks, that is, existing forms.

[0008]

[Means for Solving the Problems] The facsimile form processing method of the present invention recognizes and processes an image on a form based on the image data of the form transmitted from a facsimile, and is particularly applicable to the processing of a form transmitted from a facsimile. a first step of determining whether or not an edge mark exists in the image data; and if it is determined in the first step that an edge mark exists, the scanning accuracy of the facsimile is determined based on the position of the edge mark; a second step of calculating a correction value; a third step of storing the facsimile scanning accuracy correction value obtained in the second step in correspondence with the facsimile; and a scanning accuracy correction value stored in the third step. and a fourth step of processing the image data of the form sent from the facsimile and performing recognition processing using the facsimile.

[0009] Here, in order to store the scanning accuracy correction value for each facsimile, the scanning accuracy correction value may be stored in the third step using the telephone number of the facsimile as a key.

[0010] Furthermore, the scanning accuracy correction value stored in the third step is updated as appropriate.

[0011]

[Operation] According to the present invention, by registering the scanning accuracy correction value once, characters, tables, figures, etc. can be recognized by applying the registered scanning accuracy correction value thereafter. need only be used during registration, and after registration, normal forms without edge marks can be used.

[0012] Therefore, for example, by registering the variation in scanning accuracy between facsimile machines as a scanning accuracy correction value using a telephone number as a key, it becomes possible to use a large number of facsimile machines.

Furthermore, by appropriately updating the scanning accuracy correction value, it is possible to cope with changes in facsimile machines over time.

[0014]

[Embodiment] Next, an embodiment of the present invention will be explained with reference to FIG. In Figure 1, 3-1 to 3-N are facsimiles that scan characters, tables, figures, etc. written on forms 1 and 2 and transmit image data, 4 is a public network/ISDN network, and 5 is a facsimile data communication control 6 is a control device, and 7 is a recognition device.

The operation of this embodiment is as follows. First, in order to register the variations in scanning accuracy among facsimile machines 3-1 to 3-N, we used form 1 with pre-printed edge marks 1a for measuring scanning accuracy, and After reading, the image data is output to the public network/ISDN network 4. This image data is
Public network/ISD with facsimile phone number 3-Ka
The data is input from the N network 4 to the facsimile data communication control device 5, via the control device 6, and finally input to the recognition device 7.

Next, the recognition device 7 first searches for four edge marks 1a based on the above image data, and if there is an edge mark 1a, it recognizes the registration mode, and measures and corrects the scanning accuracy. Calculate the value and register the calculated correction value.

Details of the process from scanning accuracy measurement to correction value registration will be explained below. First, scanning accuracy is measured by finding the centers of each of the four edge marks 1a, and then calculating the number of white dots Hwck between the centers of the two edge marks 1a provided on the upper side of the form 1 (see Figure 2). . Next, form 1
The number of white dots Vwck between the centers of the two edge marks 1a provided on the left side is calculated (see FIG. 2). The correction value is calculated using the following equations (1) and (2).

NHwck=Hwck/Hwc...(
1) NVwck=Vwck/Vwc...
(2) Here, Hwc: standard value of the number of white dots in the horizontal direction Vwc: standard value of the number of white dots in the vertical direction NHwck: horizontal correction value NVwck: vertical correction value. These correction values NHwc and NVwc are registered in the recording unit in the recognition device 7 using the facsimile telephone number as a key. This completes the scanning accuracy correction value registration for the facsimile 3-K.

After the scanning accuracy registration is completed, in the case of character recognition by OCR, the above correction values NHwck and NVwck are used to recognize the characters written in Form 1. At this time, the character cutting position is corrected and character recognition is performed. do. Furthermore, in the case of a form recognition system, the pitch between the lines recognized and the frame size are corrected using correction values NHwck and NVwck. The above scanning accuracy correction value registration process is performed by facsimile 3-1 to
Perform all steps 3-N in advance.

Next, a description will be given of processing when a normal form 2 without edge marks is transmitted from the facsimile 3-K. The image data of the normal form 2 is input together with the telephone number 3-Ka from the public network/ISDN network 4 to the facsimile data communication control device 5, and is input to the recognition device 7 via the control device 6. The recognition device 7 searches for the edge mark 1a based on the input image data. In this case, since there is no edge mark 1a, the recognition device 7 recognizes the mode as the normal mode and skips the above-described scanning accuracy correction value registration process. Next, the recognition device 7 checks the telephone number of the facsimile 3-K that transmitted the image data, searches for the correction value of the facsimile 3-K from the telephone number 3-Ka, and sets the correction values NHwck, NVw.
Take out ck. As a result, the correction value NHwck,N
OCR recognition and form recognition are performed using Vwck. Here, the actual correction is performed by multiplying the finally determined number of dots by a correction value.

Note that the quality of correction can be improved by registering the correction values NHwck and NVwck from time to time, such as during regular inspections, taking into account the aging of the facsimile machine. It is also necessary when replacing facsimile machines. Note that in this embodiment, the correction value registration process is as follows:
Although the recognition device 7 determines the presence or absence of the edge mark 1a and automatically performs the registration process, it is also possible to instruct whether or not to perform the registration process using a command or the like from the control device 6.

As is clear from the above description, according to the above embodiment, the scanning accuracy of the facsimile used can be corrected even when forms without edge marks are used.
This has the effect of enabling highly accurate recognition in OCR recognition/form recognition systems. In particular, in a form recognition system, this is extremely effective in recognizing the format of already printed forms (of course, no edge marks are printed).

[0023]

[Effects of the Invention] As is clear from the above explanation, according to the present invention, it is possible to perform highly accurate recognition processing even for forms that are not provided with edge marks for correcting facsimile scanning accuracy. Become.

[Brief explanation of drawings]

[Fig. 1] Block diagram showing one embodiment of the present invention

[Figure 2] Diagram showing an example of a form with edge marks

[Figure 3] Diagram showing an example of a normal form without edge marks

[Explanation of symbols]

1, 2...Form, 3-1 to 3-N...Facsimile, 4...Public network/ISDN network, 5...Facsimile data communication control device, 6...Control device, 7...Recognition device.

Claims (1)

[Claims] Claim 1: In a facsimile form processing method that recognizes and processes an image on a form based on the image data of the form transmitted from the facsimile, whether or not edge marks exist in the image data of the form transmitted from the facsimile. a first step of determining the edge mark; and if it is determined that an edge mark exists in the first step,
a second step of determining the facsimile scanning accuracy correction value based on the position of the edge mark; and a third step of storing the facsimile scanning accuracy correction value obtained in the second step in correspondence with the facsimile. and a fourth step of processing the image data of the form transmitted from the facsimile machine and performing recognition processing using the scanning accuracy correction value stored in the third step. Form processing method.