JP2000322512A - Device and method for processing document - Google Patents

Abstract

PROBLEM TO BE SOLVED: To exactly and highly accurately perform document processing by individually calculating the quantity of deviation for each area and moving the relative positions of a document image and template information just by the maximum value. SOLUTION: With an extraction block as a reference, an area overlapped with a template block is defined as an examination object area and the extraction block, in which the examination object area is greater than a prescribed value, is left as the candidate of a processing object area (a). Concerning the extraction block existent over plural template blocks, among plural areas generated by overlapping, the maximum area is set as the examination object area (b). The deviation of a central coordinate between the selected extraction block and the template block paired with the extraction block is calculated concerning all pair blocks, all the template blocks are moved just by the calculated representative deviation quantity and an area to be overlapped after this movement is defined as the processing object area.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a form processing apparatus and a form processing method, for example, a form processing apparatus and a form for aligning an input form image with image information of a template to be compared and stored in advance. Regarding the processing method.

[0002]

2. Description of the Related Art In the field of a form processing apparatus capable of automatically recognizing a large number of forms of the same type in a lump, for example, when performing a character recognition process on a form image read by a scanner or the like, it is necessary to store the form image in advance. Template information,
That is, information preliminarily registered as to what kind of character recognition processing should be performed in partial units of the template form to be compared is superimposed on the input form image, and the form image is subjected to template information. In general, a specified character recognition process is performed.

In performing such a form processing, it is necessary that the form image and the template information are accurately superimposed. However, in a general image input apparatus such as a scanner, the same form is processed. Even if it is read, there is a problem that the reading position is slightly different for each sheet due to the physical accuracy of the scanner. For this reason, if the template information is simply superimposed on the form image, a deviation occurs between the position of the template form image and the position of the input form image inside the form processing apparatus. Even if character recognition processing is performed on the form image, an accurate recognition result cannot be obtained.

Therefore, in order to prevent such a problem of misalignment, a conventional form processing apparatus employs a scanner having high transport accuracy for reading a form, or uses a scanner for positioning the form itself to be read. In addition to printing the marking of the mark, and adding the position information of the marking for alignment to the template information registered in advance, when superimposing the actually input form image on the template information, As a mark, a method of compensating for a displacement or the like is adopted.

[0005]

However, in the above-described conventional alignment method, there are great restrictions on the form design, and the processing is based on template information. Since the influence is large, it is a major factor that hinders efficient form processing.

Accordingly, an object of the present invention is to provide a form processing apparatus and a form processing method for performing form processing accurately and with high accuracy.

[0007]

To achieve the above object, a form processing apparatus according to the present invention has the following configuration.

That is, a form processing apparatus for performing a predetermined process on an input form image in accordance with position information for each area to be included in a form to be processed and template information including predetermined control instruction information. Area identification means for identifying an area included in the form image in accordance with the feature amount of the form image thus obtained, and for each area identified by the area identification means, for each of the areas stored in advance as the template information By overlaying the location information of
A shift amount calculation that detects an area corresponding to each area identified by the area identification unit based on position information of each area included in the template information, and individually calculates a shift amount from the detected area. Means for determining a processing target area for the form image by moving a relative position between the form image and the template information by the maximum value of the amount of shift calculated by the shift amount calculating means; It is characterized by having.

In addition, for example, the shift amount calculating means may calculate an overlapping portion generated by overlapping each area identified by the area identifying means with position information of each area previously stored as the template information. Detecting a corresponding region based on the area;
More preferably, when there is a plurality of areas of the overlapping portion with respect to a certain region identified by the region identifying means, the largest overlapping portion of the position information for each region included in the template information is referred to as the largest overlapping portion. It is preferable that information on an area formed together with the area is detected as the corresponding area.

Further, for example, when the number of each area identified by the area identifying means is smaller than the number of position information for each area included in the template information, the determining means cannot detect the corresponding area. The position information of the area included in the template information may be moved according to the shift amount, and a portion where the moved area overlaps the form image may be determined as the processing target area.

Alternatively, in order to achieve the above object, a form processing method according to the present invention has the following configuration.

That is, a form processing method for performing a predetermined process on an input form image in accordance with position information of each region to be included in a form to be processed and template information including predetermined control instruction information. An area identifying step of identifying an area included in the form image in accordance with the feature amount of the input form image; and for each area identified in the area identifying step, By superimposing the position information, the region corresponding to each region identified in the region identification step is detected based on the position information for each region included in the template information, and the deviation from the detected region is detected. A shift amount calculating step of individually calculating the amount, and a relative value between the form image and the template information by a maximum value of the shift amount calculated in the shift amount calculating step. Position by moving the, and having a determination step of determining a processing target area for the form image.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a form retrieval apparatus according to the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram showing a schematic configuration of a form retrieval apparatus according to an embodiment of the present invention.

In FIG. 1, an image (a form image in the present embodiment) read by an image input unit 2 such as a scanner or a camera is subjected to a general binarization process in the unit, and then converted to a digital image. The data is sent to the processor 4 as data. The processor 4 converts the image (image data) into a table, a text, an image, or the like from the input image data by a method such as a histogram extraction method of black dots as a general image feature extraction process. Is divided for each block (region). Further, the processor (C
For example, if the extracted block is a table block, the PU) 4 further obtains a detailed structure of the target table by performing a general ruled line tracing process or the like and a count value of the number of detected tables, and If so, the character code is converted by performing a general character recognition process. The information of each area obtained in this manner is stored in the memory 10 such as a RAM and / or the storage device 8 such as a hard disk as table structure data and table number data.

When the processor 4 receives the registration information of the template form and the command of the search input from the keyboard 6, the processor 4 utilizes the table number data, the table structure data and the like stored in the storage device 8 or the like. Then, a search process described later is performed, and the search result is output to an output device 12 such as a display or a printer.

The above-described hardware configuration may be dedicated to a form processing device, or may be configured by causing a general computer to which the image input unit 2 can be connected to perform a search process described later. good.

Next, the operation of the form processing by the form processing apparatus according to the present embodiment will be described. The processing described below is realized by the processor 4 operating according to a program stored in the storage device 8 or the like in advance.

FIG. 2 is a flowchart showing the form processing by the form processing apparatus according to the present embodiment, and shows the entire flow when the processor 4 processes one form image read by the image input unit 2. .

Step S200: a routine for identifying a form image read by the image input unit 2,
For example, identification may be performed by detecting an identification mark printed in a predetermined position on a form in advance, or by analyzing the layout of the form image by general image feature amount extraction processing, You may identify whether it is a report.

Step S202: template information corresponding to the identification result of the form image in step S200,
That is, information including a processing target block (hereinafter, template block) indicating what processing (for example, character recognition or barcode reading, etc.) is to be performed on which part of one form image currently being processed. Is loaded from the storage device 8 into the memory 10. Here, in the present embodiment, since there are a large number of forms to be processed and one type of form, template information registered in the storage device 8 in advance is also one type of data relating to the form. Therefore, when there are a plurality of types of input form images, the template information to be registered in the storage device 8 in advance should also be of a plurality of types. It is necessary to perform a process of determining whether to correspond to the template information in this step.

Step S204: This is a routine for performing area identification (area division) of a form image to be processed at present, and may employ, for example, a block selection technique described in Japanese Patent Application Laid-Open No. 6-068301. In the present embodiment, based on the block selection technology,
According to the form image input in step S200, block regions having the same attribute included in the image are extracted, and region identification information such as the attribute, size, and position of each extracted block region is specified.

Step S206: The template information fetched into the memory 10 in step S202 and the
Based on the area identification information extracted in step 204, a processing target area to be subjected to a predetermined process (in this embodiment, a character recognition process) is finally determined from the form image currently being processed (details are described in detail below). See below).

Step S208: In this embodiment, a general character recognition process is performed as a predetermined process for the processing target area determined in step S206.

Next, details of the processing in step S206 of FIG. 2 will be described.

FIG. 3 is a flowchart showing a process of determining a processing target area in the form processing performed by the form processing apparatus according to the present embodiment, and shows details of step S206 in FIG.

Step S300: The template block included in the template information read in step S202 is superimposed on the block (hereinafter, extracted block) extracted from the form image currently being processed by the area identification processing in step S204. In addition, a block selection routine for selecting the extracted block as a candidate for a processing target area when the area where both blocks overlap by the superposition has an area equal to or larger than a predetermined value.

Here, the block selection rule in step S300 will be described with reference to FIG. The basic selection procedure in this block selection rule is as follows: (1) With respect to the extracted block, a region overlapping with the template block is set as an inspection target region. ,
It is left as a candidate for the processing target area (see FIG. 4A).

(2) As shown in FIG. 4 (b), with respect to an extracted block existing over a plurality of template blocks, of the plurality of regions formed by overlapping the plurality of template blocks and the extracted block, The area having the largest area is set as the inspection target area.
Then, the extracted block in which the area of the set inspection target area is equal to or larger than a predetermined value is left as a candidate of the processing target area.

(3) As a pair block to be used later for calculating the amount of displacement between the template block and the extracted block, the inspection target area noted in the above procedure (1) or (2) is shared. , An extraction block and a template block are set.

Step S302: The deviation of the center coordinates between the extracted block selected in step S300 and the template block forming a pair with the extracted block is calculated for all the paired blocks, and the histogram is calculated.

Step S304: The representative deviation amount is calculated. Here, the representative shift amount is a value (vector value) having a magnitude and a direction. If the shift amount histogram calculated in step S302 includes only one peak, the representative shift amount is the peak value. In the case of, a value indicating the maximum value is set.

Step S306: It is determined whether or not the number of extracted blocks selected in step S300 is the same as the number of template blocks (if it is large, it is impossible). If the number of both blocks is the same, the process proceeds to step S308 to extract When the number of blocks is smaller than the number of template blocks, the process proceeds to step S310.

Step S308: Since the number of extracted blocks and the number of template blocks selected in step S301 are the same, all template blocks are moved by the representative deviation calculated in step S304, and after this movement, the form image is replaced with the template block of the template. Is finally determined as a processing target area.

Step S310: Since the number of extracted blocks selected in step S301 is smaller than the number of template blocks, the template blocks forming a pair block are first moved by the representative deviation calculated in step S304, and after this movement, the form image Finally, an area overlapping with the template block of the template is determined as a processing target area, and the remaining template blocks are dealt with in step S312.

Step S312: For the remaining template blocks (isolated templates) for which the processing target area could not be determined in step S310, as shown in FIG. 5, according to the size and direction represented by the representative displacement calculated in step S304. The isolated template block having no pair block is moved, and the area after the movement is set as the processing target area.

As described above, according to the present embodiment, highly accurate form alignment can be realized without taking into account marking for absorbing form deviation and layout restrictions.
Further, since the processing is performed based on the area recognition result performed for each input form, the form itself is resistant to deformation and the efficiency of the form processing can be improved.

<First Modification> In the above-described embodiment, when a plurality of template blocks overlap with an extraction block, the template block having the largest area of the overlapping portion is set as a pair block. However, the present invention is not limited to this. For example, a template block shape (for example, an aspect ratio) that is similar to the shape of an extraction block may be used as a pair block.
Alternatively, as in the above-described embodiment, a pair block may be set using an evaluation score that takes into account both the area of the portion where the template block and the extracted block overlap and the approximation of the shape.

<Second Modification> In the above-described embodiment, when the number of remaining blocks and the number of template blocks are the same in step S310 of FIG. 3, the template block is moved by the representative deviation amount, and The area after the movement is set as the processing target area. However, the present invention is not limited to this. For example, when the number of blocks is the same and only one peak of the shift amount histogram exists, the extracted block is set as the processing target area, If there is one or more peaks in the amount histogram, the extracted block is set as the processing target region for the template block indicating the representative deviation amount, and otherwise, the region in which the template block is moved by the representative deviation amount is set as the processing target region. Is also good.

[0040]

Further, another object of the present invention is to provide a storage medium (or a recording medium) in which a program code of software for realizing the functions of the above-described embodiments is recorded to a system or an apparatus, and the system or the apparatus is provided. It is needless to say that the present invention can also be achieved by a computer (or CPU or MPU) reading and executing the program code stored in the storage medium. In this case, the program code itself read from the storage medium implements the functions of the above-described embodiment, and the storage medium storing the program code constitutes the present invention. The functions of the above-described embodiments are not only realized by the computer executing the readout program code, but also the operating system running on the computer based on the instructions of the program code.
This includes the case where the (OS) performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

Further, after the program code read from the storage medium is written into a memory provided in a function expansion card inserted into the computer or a function expansion unit connected to the computer, the program code is read based on the instruction of the program code. This also includes the case where the CPU provided in the function expansion card or the function expansion unit performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

[0042]

As described above, according to the present invention,
Provided is a form processing apparatus and a form processing method for performing form processing with accuracy and high accuracy.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of a form search device according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating a form process performed by the form processing apparatus according to the embodiment.

FIG. 3 is a flowchart showing a process of determining a processing target area in form processing performed by the form processing apparatus according to the embodiment.

FIG. 4 is a diagram illustrating a rule of block selection.

FIG. 5 is a diagram illustrating a case where a processing target area is determined for an isolated template.

[Explanation of symbols]

 2: image input unit, 4: processor, 6: keyboard, 8: storage device, 10: memory, 12: output unit,

Claims (9)

[Claims] 1. A form processing apparatus for performing predetermined processing on an input form image according to position information for each area to be included in a form to be processed and template information including predetermined control instruction information, Area identification means for identifying an area included in the form image according to the feature amount of the created form image, and for each area identified by the area identification means,
By superimposing the position information of each of the areas stored in advance as the template information, the area corresponding to each of the areas identified by the area identifying means is replaced with the position information of each of the areas included in the template information. And a displacement amount calculating means for individually calculating a displacement amount with respect to the detected area, and a relative value between the form image and the template information by a maximum value of the displacement amount calculated by the displacement amount calculating means. Determining means for determining a processing target area for the form image by moving a position of the form image. 2. The method according to claim 1, wherein the shift amount calculating unit is configured to overlap an area identified by the area identifying unit with position information of each area stored in advance as the template information. 2. A corresponding area is detected based on the following.
Form processing device as described. 3. When there is a plurality of areas of the overlapping portion with respect to a certain area identified by the area identifying means, the shift amount calculating means calculates position information of each area included in the template information. 3. The form processing apparatus according to claim 2, wherein information of an area forming a largest overlapping portion together with the area is detected as the corresponding area. 4. The template which cannot detect the corresponding area when the number of each area identified by the area identifying means is smaller than the number of position information for each area included in the template information. 2. The form processing according to claim 1, wherein position information of an area included in the information is moved according to the shift amount, and a portion where the moved area overlaps with the form image is determined as the processing target area. apparatus. 5. A character recognizing means for performing character recognition on each area identified by the area identifying means according to predetermined control instruction information included in the template information after the positioning by the positioning means. The form processing apparatus according to claim 1, further comprising: 6. A form processing method for performing predetermined processing on an input form image in accordance with position information of each area to be included in a form to be processed registered in advance and template information including predetermined control instruction information. An area identifying step of identifying an area included in the form image in accordance with the feature amount of the input form image; and for each of the areas identified in the area identifying step, for each of the areas included in the template information. By superimposing the position information, the region corresponding to each region identified in the region identification step is detected based on the position information for each region included in the template information, and the deviation from the detected region is detected. A shift amount calculating step of individually calculating the amount, and only the maximum value of the shift amount calculated in the shift amount calculating step,
Determining a region to be processed for the form image by moving a relative position between the form image and the template information. 7. In the shift amount calculating step, each area identified in the area identifying step is overlapped with position information of each area included in the template information, based on an area of an overlapping portion generated by overlapping each area. 7. The form processing method according to claim 6, wherein a corresponding area is detected. 8. In the shift amount calculating step, when there is a plurality of areas of the overlapping portion with respect to a certain area identified in the area identifying step, the position information of each area included in the template information is obtained. 8. The form processing method according to claim 7, wherein information of an area forming a largest overlapping portion together with the area is detected as the corresponding area. 9. In the determining step, when the number of each area identified in the area identifying step is smaller than the number of pieces of position information for each area included in the template information, the corresponding template cannot be detected. 7. The form processing according to claim 6, wherein the position information of the area included in the information is moved according to the shift amount, and a portion where the moved area overlaps with the form image is determined as the processing target area. Method.