JP4842872B2 - Form processing device - Google Patents

Description

  The present invention relates to a form processing apparatus that reads contents entered in a form and converts them into data, and in particular, it is possible to automatically detect entry contents and convert it into data without setting position information of the entry area of the form in advance. The present invention relates to a form processing apparatus.

  2. Description of the Related Art A form processing apparatus is known that reads a form such as a questionnaire survey response form or various application forms and converts the written contents into data. In such a form processing apparatus, format information indicating the coordinate position of each entry column is stored in advance as position information corresponding to the format of the form. The form processing apparatus specifies the area of each entry field in the completed form based on the position information, performs a reading process on each specified area, and converts the entry contents into data.

In the technology disclosed in Patent Document 1 below, in correspondence with the format of the form,
Layout information indicating the coordinate position of each entry field is preliminarily embedded in the form as position information, and based on this position information, the completed form is read, and the contents of the entry are input.
JP 2005-157536 A

  However, in the above prior art, it is necessary to manually set position information for each different form. Further, when an entry field is added or the layout of the entry field is changed and the format of the form is changed, the setting operation has to be performed again, which is troublesome.

  Therefore, there has been a demand for a form processing apparatus that can automatically detect entry information from a completed form and convert it into data without setting position information in advance.

  The present invention adopts the following configuration in order to solve the above points.

The form processing apparatus according to the present invention divides an original form image into a plurality of blocks, acquires each feature data corresponding to the position of each block , and generates a watermark image generated based on the feature data. A form processing device for reading the contents of a form that is embedded in an image and printing it, and an image acquisition unit that reads a form that has been filled in to obtain an overall image, and a watermark image from the acquired overall image And a separate acquisition unit that separates and acquires the entry form image including the entry contents and the original form image, and an analysis acquisition unit that analyzes the acquired watermark image and acquires a plurality of feature data as original feature data, respectively. When, collected each feature data corresponding to the position of each block from the obtained input form image as the current characteristic data, respectively A characteristic acquisition unit which, Bei example a determination unit for determining whether the entry contents by comparing each original feature data and each current characteristic data, the density of each pixel constituting the watermark image is a predetermined density threshold Based on the original form image and the gradation density different from the density of each pixel constituting the entry content, the separation acquisition unit obtains density data indicating the density of each pixel constituting the entire image, and obtains a predetermined density A watermark image and a form image are separated based on a threshold value .

  According to the form processing apparatus of the present invention, the watermark image and the form image are separated from the entire image of the form in which the entry contents are entered, the feature data obtained by analyzing the watermark image, and the form image Since the presence / absence of the entry contents is determined by comparing with the feature data acquired from the entry data, the entry contents can be automatically detected without setting the position information of the entry area in advance. This eliminates the need for human setting work for position information. Therefore, even if the format of the form is changed, a form processing apparatus that can automatically detect the contents of entry and convert it into a data is realized.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram illustrating a configuration of a form processing apparatus according to the first embodiment of the present invention.
Using this form processing apparatus 10, the entry content can be detected and read from the form.

  As illustrated in FIG. 1, the form processing apparatus 10 according to the present exemplary embodiment includes an image acquisition unit 11, a separation acquisition unit 12, a pixel number acquisition unit 13, an analysis acquisition unit 14, a determination unit 15, an additional write content storage unit 16, and a region. A generation unit 17, a noise detection unit 18, an analysis extraction unit 19, a temporary storage unit 20, an output unit 21, and a control unit 22 are provided.

  The image acquisition unit 11 reads a form and acquires an image of the form as an entire image. In the present embodiment, it is assumed that the image acquisition unit 11 acquires a 256-level multi-value gray image as the entire image. The acquired entire image is stored in the temporary storage unit 20.

FIG. 2 is a diagram illustrating an example of a form.
2A shows an image of a form format, that is, an original form image 31, and FIG. 2B shows an original form 32 printed by embedding a watermark image in the original form image 31. FIG. . The image acquisition unit 11 reads the form in which the entry contents are entered, and acquires the entire image 33 as shown in FIG.

  The separation / acquisition unit 12 reads the entire image 33 from the temporary storage unit 20 and separates and acquires the entry form image including the entry content and the original form image 31 and the watermark image from the entire image 33.

  On the paper surface of the form 32 shown in FIG. 2B, a watermark image having a light gradation density is embedded and printed so as to be separated from the original form image 31 in order to enable separation from the original form image 31. Has been. That is, the density of the watermark image is set lower than the predetermined density threshold in this embodiment. Further, in order to enable separation from the watermark image, the entry means for the entry content is restricted so that the gradation density of the entry content becomes deeper than the predetermined density threshold described above. To do. The separation acquisition unit 12 stores the above-described predetermined density threshold in advance, and separates the entry form image and the watermark image based on the density threshold. In the present embodiment, “5” is stored in the separation acquisition unit 12 as the density threshold value.

Hereinafter, the procedure of the separation process of the entry form image and the watermark image by the separation acquisition unit 12 will be briefly described with reference to FIG.
FIG. 3 is an explanatory diagram of the separation process of the entry form image and the watermark image.

The separation acquisition unit 12 first acquires the density data of each pixel constituting the entire image 33 (FIG. 2C), that is, the pixel density value, and uses this pixel density value as coordinates for specifying the position of the pixel. The data is stored in the temporary storage unit 20 in association with the data.
FIG. 3A shows an example of each pixel density value acquired from the entire image 33 by the separation acquisition unit 12. For example, the separation acquisition unit 12 acquires a pixel density value “2” for the pixel 35 a in the entire image 33. Further, the separation acquisition unit 12 acquires a pixel density value “6” for the pixel 36a.

  Subsequently, the separation acquisition unit 12 compares the acquired pixel density value with the stored density threshold value for each pixel of the entire image 33, and specifies a pixel whose pixel density value is less than the density threshold value. Then, the separation acquisition unit 12 sets the pixel density value of each specified pixel as the pixel density value of the watermark image. Further, the pixel density value of each pixel whose pixel density value is equal to or higher than the density threshold value is set as the pixel density value of the entry form image. For example, since the pixel density value “2” of the pixel 35a is less than the density threshold value “5”, the pixel density value “6” of the pixel 36a is equal to or higher than the density threshold value “5”. Therefore, the pixel density value of the entry form image is used. In this way, the separation acquisition unit 12 separates the watermark image and the entry form image from the entire image. Each separated pixel density value is stored in the temporary storage unit 20 in association with position data indicating the position of the pixel.

  FIG. 3B and FIG. 3C show the pixel density values of the watermark image and the entry form image separated from the entire image. As described above, in the entire image (FIG. 3A), the pixel 35a having the pixel density value “2” is separated as a watermark image. Therefore, in the watermark image shown in FIG. The pixel density value of the pixel 35b is “2”. In the entry form image shown in FIG. 3C, the pixel density value of the pixel 35c is “0”. Further, in the entire image (FIG. 3A), the pixel 36a having a pixel density value of “6” is separated as an entry form image. Therefore, in the entry form image FIG. 3C, the pixel 36c. The density value is “6”, and in the watermark image (FIG. 3B), the pixel density value of the pixel 36b is “0”.

  FIG. 2D shows an entry form image 34 obtained by being separated and obtained by the separation obtaining unit 12. The entry form image 34 from which the watermark image has been removed from the entire image 33 (FIG. 2C) includes the entry contents and the original form image 31 as shown in FIG.

As the feature acquisition unit, the pixel number acquisition unit 13 reads the entry form image 34 from the temporary storage unit 20 and acquires the current number of pixels as feature data from the entry form image 34.
FIG. 4 is an explanatory diagram of the current pixel number acquisition process in the entry form image.

  First, the pixel number acquisition unit 13 divides the entry form image 34 (FIG. 2D) into a grid having a predetermined size to obtain a plurality of blocks, as shown in FIG. In FIG. 4, as an example, the pixel number acquisition unit 13 divides the entry form image 34 so that each block has a size of 3 pixels wide × 3 pixels high. Based on the control of the control unit 22, the pixel number acquisition unit 13 divides the entry form image 34 into blocks of a predetermined size. Each divided block is given block identification data for identifying the position of the block. In this embodiment, each block has data (x) indicating the number of blocks from the origin position in the coordinate system in which the upper left in the entry form image 34 is the origin position, the left end of the form is the x axis, and the upper end is the y axis. , Y) is given as block identification data. For example, in FIG. 4B, block identification data (1, 1) is given to the block 37b. Further, block identification data (1, 2) is given to the block 38b.

  Subsequently, the pixel number acquisition unit 13 acquires the number of black pixels included in each block as feature data in the block, that is, the current pixel number, and associates it with the block identification data of the block in the temporary storage unit 20. Remember me. Here, the number of black pixels acquired by the pixel number acquisition unit 13 indicates the number of pixels whose pixel density value is not “0” among a plurality of pixels included in each block.

  FIG. 4B shows the number of current pixels acquired from each block of the entry form image. For example, in the entry form image shown in FIG. 4A, the pixel density values of the pixels included in the upper left block 37a are all “0”. As shown in b), “0” is acquired as the current pixel number of the block 37b. In addition, since the block 38a (FIG. 4A) includes two pixels whose pixel density value is not “0”, as shown in FIG. 4B, the current number of pixels in the block 38b is “ 2 "is acquired.

  The analysis acquisition unit 14 reads a watermark image from the temporary storage unit 20 and performs an analysis process on the watermark image.

  The watermark image embedded in the original form image 31 is generated based on the number of original pixels as the characteristic data of the original form image 31. This number of original pixels corresponds to the number of black pixels in each block obtained by dividing the original form image 31 into a grid of a predetermined size. For example, the original form image 31 is divided into blocks each having a size of 3 pixels wide by 3 pixels high, and the number of original pixels is acquired for each block together with block identification data for identifying the block. In this embodiment, each block includes data indicating the number of blocks from the origin position in the coordinate system in which the upper left portion of the original form image 31 is the origin position, the left end portion of the original form image is the x axis, and the upper end portion is the y axis. Is given as block identification data.

  Each block identification data and the corresponding number of original pixels are expressed by a watermark image as characteristic information of the original form image 31 together with the size of each block, that is, the division size of the original form image 31. The watermark image is composed of a plurality of dot patterns, and each dot pattern represents a different signal depending on the dot arrangement. The watermark image expresses the above-described feature information by arranging these signals, that is, dot patterns in accordance with a certain rule.

  The analysis acquisition unit 14 reads out each dot pattern by performing a filtering process on the watermark image generated in this way. Then, the feature information represented by the watermark image is acquired based on the arrangement of the read dot patterns. Each block identification data and the number of original pixels of the block in the original form image 31 included in the acquired feature information are stored in the temporary storage unit 20, and the division size is notified to the control unit 22.

The determination unit 15 determines the presence / absence of entry contents in each block of the entry form image 34 based on the current pixel number and the original pixel number.
The determination unit 15 reads out the current pixel number and the original pixel number in the same block from the temporary storage unit 20 based on the block identification data given to each block, and the number of pixels as difference data indicating the difference between them. Calculate the difference. This pixel number difference indicates the difference between the current pixel number acquired from the entry form image 34 including the entry contents and the original form image 31 and the original number of pixels acquired from the original form image 31. This is considered to correspond to the number of black pixels in the content. The determination unit 15 stores in advance a pixel number difference threshold value as a predetermined difference threshold value, and compares the calculated pixel number difference with the stored pixel number difference threshold value, thereby Determine presence or absence. That is, when the pixel number difference is equal to or larger than the pixel number difference threshold, it is determined that the block is a filled-in block having written contents. In this embodiment, it is assumed that “6” is stored in the determination unit 15 as the pixel number difference threshold.

  In addition, the determination unit 15 determines whether the entry content is newly added content that has been newly entered, or whether the content has been altered. When the original pixel number acquired from the original form image 31 is “0” in each entry block, the determination unit 15 determines that the entry content in the entry block is newly added content, and the original pixel number Is not “0”, it is determined that the entry contents in the entry block are falsification contents obtained by falsifying the format. Then, the determination unit 15 determines that the entry-added block having the newly added content is the additionally-added block, and notifies the control unit 22 of the block identification data of the additionally-added block. In addition, a block with entry having falsification content is determined to be a block with falsification, and the determination result is notified to the control unit 22.

FIG. 5 is a diagram (part 1) illustrating an example of an entry form image.
In the entry form image 34 shown in FIG. 5, since the number of original pixels in the entry form image 31 (FIG. 2A) is not “0” in the entry blocks 40-1 and 40-2, the determination unit 15 Then, it is determined that the block is a falsified block including falsified content whose format has been falsified.

  The additional write content storage unit 16 is a storage unit that stores block identification data of a block that is determined by the determination unit 15 to be a block with additional write.

  The area generation unit 17 generates an additional write area in the entry form image 34. The area generation unit 17 refers to the additional writing content storage unit 16 and determines the presence / absence of a block with additional writing in the entry form image 34 based on whether the block identification data is stored. When there is a block with additional recording, that is, when block identification data is stored in the additional recording content storage unit 16, the area generation unit 17 generates each adjacent block with additional recording as an additional recording area based on the block identification data. Then, an area identification number is assigned to the generated additional recording area, position data indicating the position of the additional recording area is acquired, and the control unit 22 is notified of the area identification number and position data. When there is no block with additional writing, that is, when block identification data is not stored in the additional writing content storage unit 16, the region generation unit 17 notifies the control unit 22 of the determination result.

  For example, the area generation unit 17 generates the additional write areas 39-1 to 7 from each additional write block in the entry form image 34 (FIG. 5). And the area | region production | generation part 17 assign | provides area | region identification numbers 1-7 to each additional recording area | region 39-1-7. Further, the area generation unit 17 obtains the maximum and minimum x coordinate data and y coordinate data on the contour line surrounding each additional write area 39-1 to 7 based on the block identification data stored in the additional write content storage section 16. , Get each. Each acquired coordinate data is stored in the additional write content storage unit 16 as position data indicating the position of the additional write area in association with the area identification number.

  The noise detection unit 18 detects the area of the additional recording region, that is, the region size, for each additional recording region generated by the region generation unit 17. In the present embodiment, the noise detection unit 18 detects the number of blocks included in each additional write area as the area size. Then, the noise detecting unit 18 determines whether or not the additional recording area is a noise area based on the detected area size. The noise detection unit 18 stores a predetermined size threshold in advance, and when the detected area size is equal to or smaller than the stored size threshold, the additional recording area is a noise area to which dust or dirt is attached. judge. The area determined to be a noise area is excluded from the additional recording area. In the present embodiment, it is assumed that “4” is stored in the noise detection unit 18 as the size threshold value.

  For example, since the area size of the additional recording area 39-1 is “3” among the plurality of additional recording areas 39-1 to 7 in the entry form image 34 (FIG. 5), the noise detecting unit 18 It is determined that 39-1 is a noise region. As a result, the additional recording area 39-1 is excluded from the additional recording area.

  The analysis extraction unit 19 performs an analysis process on the entry form image 34 and extracts entry contents as newly added contents from the additional entry areas 39-2 to 7 that are not noise areas. The analysis extraction unit 19 performs OCR (Optical Character Reading) processing, image cut-out processing, and the like, extracts newly added content, and stores it in the additional write content storage unit 16 in association with the area identification number.

  The temporary storage unit 20 is a storage unit that temporarily stores various data generated during processing.

  The output unit 21 outputs each newly added content extracted by the analysis extracting unit 19 and stored in the additional write content storage unit 16. The output unit 21 reads out the newly added content and the corresponding position data included in each of the additional writing areas 39-2 to 39-7 from the additional writing content storage unit 16, and outputs them. The output data output by the output unit 21 is input and stored in a database (not shown), for example.

  The control unit 22 controls the above-described units 11 to 21.

Next, the operation of the form processing apparatus 10 according to the present embodiment will be described with reference to the drawings. Here, a flow of processing for reading a form corresponding to the entry form image 34 shown in FIG. 5 and outputting entry contents will be described.
FIG. 6 is a flowchart (part 1) showing the operation in the first embodiment of the form processing apparatus according to the present invention, and FIG. 7 is a flowchart (part 2) showing the operation in the first embodiment of the form processing apparatus according to the present invention. ).

  When the form processing is started in the form processing apparatus 10, the image acquisition unit 11 first reads the form and acquires the entire image 33 (FIG. 2C) under the control of the control unit 22 (step S101). The acquired entire image 33 is stored in the temporary storage unit 20.

  Subsequently, the control unit 22 controls the separation acquisition unit 12 to execute the separation process of the watermark image and the form image 34. The separation acquisition unit 12 reads the entire image 33 (FIG. 2C) from the temporary storage unit 20, acquires the pixel density value of each pixel constituting the entire image 33, and the pixel density value is less than the density threshold. Each pixel is separated as a watermark image and each pixel having a pixel density value equal to or higher than the density threshold value as an entry form image (step S102). The separated watermark image and entry form image 34 (FIG. 2D) are stored in the temporary storage unit 20.

  Next, the control unit 22 controls the analysis acquisition unit 14 to perform a watermark image analysis process. The analysis acquisition unit 14 reads a watermark image from the temporary storage unit 20 and performs a filtering process on the watermark image. Then, the analysis acquisition unit 14 acquires the feature information of the original form image 31 included in the watermark image. The read feature information includes the division size of the original form image 31, the block identification data of each block divided from the original form image 31, and the corresponding number of original pixels. The division size is notified to the control unit 22, and the acquired block identification data and the number of original pixels are stored in the temporary storage unit 20 (step S103).

  When the notification of the division size is received, the control unit 22 controls the pixel number acquisition unit 13 to acquire the current pixel number from the entry form image 34 (FIG. 2D). The pixel number acquisition unit 13 reads the entry form image 34 from the temporary storage unit 20, divides the entry form image 34 into blocks of the division size notified from the control unit 22, and blocks identification data into these blocks. Is granted. Then, the pixel number acquisition unit 13 acquires the number of black pixels included in each block as the current pixel number (step S104). Each acquired number of current pixels is stored in the temporary storage unit 20 in association with the block identification data of the block.

  Subsequently, the control unit 22 controls the determination unit 15 to determine the presence / absence of entry contents in each block. The determination unit 15 reads the current pixel number and the original pixel number for the same block one by one from the temporary storage unit 20 based on each block identification data, and calculates the pixel number difference. Then, the determination unit 15 compares the calculated pixel number difference with the stored pixel number difference threshold value (step S105).

  When the pixel number difference is equal to or greater than the pixel number difference threshold, the determination unit 15 determines that the block is a filled-in block having written content (step S105). Then, it is determined whether or not the number of original pixels corresponding to the block is 0 (step S106).

  When the number of original pixels is 0 (step S106), the determination unit 15 determines that the entry content included in the entry block is newly added content, and also determines that the entry block is a block with additional writing. (Step S107). Then, the determination unit 15 notifies the control unit 22 of the block identification data of the block with additional writing along with the determination result.

  Upon receiving the block identification data from the determination unit 15, the control unit 22 stores the block identification data in the additional write content storage unit 16 (step S108).

  In step S106, when the number of original pixels is not 0, the determination unit 15 determines that the entry content included in the entry block is falsification content and determines that the entry block is a falsification block ( Step S109). Then, the determination unit 15 notifies the control unit 22 of the determination result.

  In step S105, when the pixel number difference is less than the pixel number difference threshold, the determination unit 15 determines that the block is an unfilled block (step S110), and notifies the control unit 22 of the determination result. .

  The control unit 22 notified of the determination result from the determination unit 15 determines whether there is an undetermined block (step S111). When there is an undetermined block, the control unit 22 causes the determination unit 15 to perform the processing from step S105 onward for the block.

  If it is determined that there is no undetermined block (step S111), the control unit 22 controls the region generation unit 17 to determine whether there is a block with additional writing (step S112). The area generation unit 17 determines the presence / absence of a block with additional writing based on the presence / absence of storage of block identification data in the additional writing content storage unit 16. If it is determined that there is no block with additional writing, the form processing apparatus 10 ends the process.

  If it is determined that there is a block with additional writing (step S112), the area generation unit 17 generates an adjacent block with additional writing as an additional writing area based on each block identification data stored in the additional writing content storage unit 16. Then, an area identification number is assigned to each generated additional recording area, and position data of each additional recording area is acquired (step S113). The area generation unit 17 generates the additional write areas 39-1 to 7 in the entry form image 34 (FIG. 5), assigns the area identification numbers 1 to 7, and the position data of the additional write areas 39-1 to 7-7. The area identification number and the corresponding position data are acquired and stored in the additional write content storage unit 16.

  Subsequently, the control unit 22 controls the noise detection unit 18 to detect the region size of the additional recording region, and compares the region size with the size threshold (step S114). First, the noise detection unit 18 detects the area size “3” of the additional recording area 39-1 to which the area identification number “1” is assigned, and compares it with the size threshold “4”. Since the detected area size is equal to or smaller than the size threshold, the noise detection unit 18 determines that the additional recording area 39-1 is a noise area (step S115). The determination result is notified to the control unit 22, and the control unit 22 excludes the additional recording area 39-1 from the additional recording area.

  Subsequently, the control unit 22 determines whether or not there is a remaining additional recording area (step S118). When there is a remaining additional recording area, the control unit 22 detects the area size by the noise detection unit 18 (step S114). Next, the noise detection unit 18 detects the area size “9” of the additional recording area 39-2 to which the area identification number “2” is assigned and compares it with the size threshold “4” (step S114). Since the detected area size is larger than the size threshold, the noise detection unit 18 determines that the additional recording area 39-2 is not a noise area, and notifies the control unit 22 of the determination result.

  Upon receiving the notification, the control unit 22 causes the analysis extraction unit 19 to perform analysis processing on the additional recording area 39-2 to extract newly added content (step S116). The analysis extraction unit 19 performs OCR processing on the additional recording area 39-2 (FIG. 5), performs character recognition of the newly added content, and recognizes and extracts the character data “18” as the newly added content. The extracted newly added content is stored in the additional write content storage unit 16 in association with the area identification number “2” (step S116).

  Then, the control unit 22 reads out and outputs the newly added content “18” and the corresponding position data from the additional write content storage unit 16 by the output unit 21 (step S117).

  Next, the control unit 22 determines whether there is a remaining additional recording area (step S118). And the same process is performed also with respect to the remaining write-once areas 39-3 to 39-7 (steps S116 to S117). The analysis extracting unit 19 recognizes and extracts the character data “2” and “27” from the additional recording areas 39-3 and 4 as newly added contents. The analysis extraction unit 19 determines that character recognition is not possible for the additional recording area 39-5, and performs image extraction processing. Then, the analysis extraction unit 19 extracts the seal image as newly added content. Further, character data “Taro Oki” and “(1)” are recognized and extracted from the additional recording areas 39-6 and 7 respectively. The newly added content extracted from each additional recording area 39-2 to 7 is stored in the additional recording content storage unit 16 in association with the region identification number (step S116), and then read out together with the corresponding position data by the output unit 21. Is output (step S117). Thereby, the form processing by the form processing apparatus 10 ends.

  As described above, the form processing apparatus according to the present embodiment separates the watermark image and the entry form image from the entire image obtained by reading the form in which the entry contents are entered, analyzes the watermark image, and analyzes the original form image. The original pixel number is acquired, and the current pixel number is acquired from the entry form image. Since the presence / absence of the entry content is determined based on the difference between the current pixel number and the original pixel number, the entry content can be automatically detected without setting the position information of the entry area in advance. This eliminates the need for human setting work for position information. Therefore, a form processing apparatus that can automatically process all forms even when the form format is changed is realized. In addition, since it is possible to automatically process all the form bundles in which forms of different formats are mixed, the applicable range is expanded. Further, since the falsification content and noise can be excluded from the detected entry content based on the number of original pixels and the area size, highly reliable data can be acquired.

  In addition, although the present Example has taken the structure which outputs only the entry content determined to be new addition content, it is not limited to this. For example, it is possible to adopt a configuration in which flag data indicating that falsification has been made is given to the entry content determined to be falsification content, and all detected entry content is output. Further, flag data may be added to the processing result of the form in which the contents of tampering are detected. Furthermore, a configuration may be added in which an entry form image including the contents of falsification is displayed on the screen to warn the operator. This will be described in detail in the second embodiment.

  In the present embodiment, the configuration is such that the entry content and the position data are output. However, it is also possible to adopt a configuration in which the content is output in combination with attribute information indicating the attribute of the entry content. As a result, it is possible to automatically read the contents of entry from a large amount of forms and convert it into data, and to perform a counting process. A method for assigning attribute information to the entry content will be described in detail in the third embodiment.

FIG. 8 is a block diagram illustrating the configuration of the form processing apparatus according to the second embodiment of the present invention.
The form processing apparatus 50 according to the second embodiment is different from the first embodiment in the configuration in which a shape data acquisition unit 51, a falsification content storage unit 54, a seal detection unit 56, and a display unit 58 are added.
In the present embodiment, the same components as those in the first embodiment are denoted by the same reference numerals, and detailed description thereof will be omitted.

  As illustrated in FIG. 8, the form processing apparatus 50 according to the present exemplary embodiment includes an image acquisition unit 11, a separation acquisition unit 12, a pixel number acquisition unit 13, a shape data acquisition unit 51, an analysis acquisition unit 52, a determination unit 53, and an additional note. The content storage unit 16, the falsification content storage unit 54, the region generation unit 55, the seal detection unit 56, the analysis extraction unit 57, the temporary storage unit 20, the output unit 21, the display unit 58, and the control unit 59 are provided.

As the feature acquisition unit, the shape data acquisition unit 51 calculates shape data as feature data using a shape extraction filter for the form image divided into blocks of a predetermined size by the pixel number acquisition unit 13. Acquired as shape data.
The shape extraction filter is a filtering process for extracting a specific shape, and is used, for example, to extract a contour line having a specific direction and width. In the present embodiment, the shape data is acquired using the technique disclosed in Japanese Patent Application Laid-Open No. 2006-157831.

  Based on the control of the control unit 59, the shape data acquisition unit 51 performs the above-described filtering process on each block obtained by dividing the form image into a grid of a predetermined size. Get the corresponding current shape data. The acquired current shape data is notified to the control unit 59 together with the block identification data of the block.

  The analysis acquisition unit 52 reads a watermark image from the temporary storage unit 20 and performs an analysis process on the watermark image. In this embodiment, the watermark image embedded in the original form image 31 (FIG. 2A) is further converted into a plurality of original shape data in addition to the original pixel number acquired from each block of the original form image 31. Has been generated based on. This original shape data corresponds to the shape data calculated using the shape extraction filter for each block obtained by dividing the original form image 31 into a grid of a predetermined size. In the present embodiment, the ledger form image 31 is divided into blocks of the same size as when the number of original pixels is acquired, and the technique disclosed in Japanese Patent Application Laid-Open No. 2006-157831 is applied to the blocks. The original shape data is acquired.

  Each original shape data acquired from each block as described above is associated with block identification data. Each block identification data and the corresponding original pixel number and original shape data are expressed by the watermark image as the feature information of the original form image 31 together with the division size of the original form image 31. The analysis acquisition unit 52 performs filtering processing on the watermark image to acquire feature information expressed by the watermark image. Each block identification data in the original form image 31 included in the acquired feature information, the number of original pixels of the block, and original shape data are stored in the temporary storage unit 20, and the division size is notified to the control unit 59.

  The determination unit 53 determines the presence / absence of entry contents in each block of the entry form image based on the current pixel number and the original pixel number. When it is determined that there is an entry content, the determination unit 53 further determines whether the entry content is newly added content or falsified content based on whether the original pixel number is “0” or not. Judgment is made. The determination unit 53 determines that the filled-in block having the newly added content is a block with additional writing, and notifies the control unit 59 of the block identification data of the block with additional writing. Further, it is determined that the filled-in block having the falsification content is a falsified block, and notifies the control unit 59 of the block identification data of the falsified block.

Further, the determination unit 53 determines whether or not there is falsification content in each block of the entry form image based on the current shape data and the original shape data.
The determination unit 53 reads the corresponding current shape data and original shape data from the temporary storage unit 20 for each block in which the pixel number difference is less than the pixel number difference threshold, and uses the shape data as difference data indicating the difference between them. Calculate the difference. This shape difference indicates a difference between the current shape data acquired from the entry form image and the original shape data acquired from the original form image 31. As described above, the difference in the number of pixels of the block is less than the threshold value of the number of pixels difference, so it is considered that the format has been altered in the block. The determination unit 53 stores a shape difference threshold value as a predetermined difference threshold value in advance, and compares the calculated shape difference with the stored shape difference threshold value to determine the presence / absence of falsification content in the block. To do. That is, when the shape difference is equal to or greater than the shape difference threshold, it is determined that the block is a tampered block having tampering content. Then, the determination unit 53 notifies the control unit 59 of the block identification data of the tampered block.

FIG. 9 is a diagram (part 2) illustrating an example of an entry form image.
In the entry form image 60 shown in FIG. 9, each block indicated by 62 has a pixel number difference from the original form image 31 (FIG. 2A) less than the pixel number difference threshold, but the shape difference is the shape difference. Above the threshold. Therefore, the determination unit 53 determines that the block is the tampered block 62 including the contents of tampering.

  The falsification content storage unit 54 is a storage unit that stores block identification data of a block that is determined to be a falsified block by the determination unit 53.

The area generation unit 55 generates an additional write area in the entry form image and notifies the control unit 59 of the area identification number and position data of each additional write area.
Further, the region generation unit 55 refers to the falsification content storage unit 54 and determines the presence or absence of a falsified block in the entry form image based on whether or not the block identification data is stored. When there is a block with falsification, that is, when block identification data is stored in the falsification content storage unit 54, the area generation unit 55 generates each adjacent falsified block as a falsification area based on the block identification data. Then, an area identification number is assigned to the generated falsified area, position data indicating the position of the falsified area is acquired, and the area identification number and position data are notified to the control unit 59. When there is no falsified block, that is, when block identification data is not stored in the falsification content storage unit 54, the region generation unit 55 notifies the control unit 59 of the determination result.

  For example, in the entry form image 60 (FIG. 9), the area generation unit 55 assigns the area identification number 1 to each falsified block 62 as a falsified area.

  The seal stamp detection unit 56 stores a predetermined size threshold value and a predetermined circularity threshold value, and detects a seal stamp area on which the seal stamp is imprinted from each additional write area generated in the entry form image. The seal stamp detection unit 56 first detects the area of the additional recording region, that is, the region size, for each additional recording region generated by the region generation unit 55. In this embodiment, the seal stamp detection unit 56 detects the number of blocks included in each additional write area as the area size. Then, the seal stamp detection unit 56 compares the detected area size with the stored size threshold value. If the area size is less than the size threshold, it is determined that the additional recording area is not a seal area.

に基づき計算する。ここで、πは円周率、Ｓは追記領域の領域サイズであり、ｌは追記領域の周囲の長さである。印鑑検出部５６は、計算した円形度ｅを記憶している円形度閾値と比較し、円形度ｅが円形度閾値以上である場合、該領域が印鑑領域であると判定する。本実施例では、印鑑検出部５６は、サイズ閾値として「１６」、円形度閾値として「０．５」を記憶しているものとする。 The seal stamp detection unit 56 further calculates the circularity e of the additional recording area with respect to the additional recording area whose area size is equal to or larger than the size threshold.

Calculate based on Here, π is the circumference ratio, S is the area size of the additional recording area, and l is the circumference of the additional recording area. The seal stamp detection unit 56 compares the calculated circularity e with a stored circularity threshold, and when the circularity e is equal to or greater than the circularity threshold, it determines that the region is a seal stamp region. In this embodiment, the seal stamp detection unit 56 stores “16” as the size threshold and “0.5” as the circularity threshold.

For example, the input form image 60 (FIG. 9), the area size of the additional write area 61 - 4 32, circularity is 0.74. Therefore, seal detecting unit 56 determines the additional write area 61- 4 as a seal area.

  In addition, for the detection of the seal region, since the shape of the seal is not limited to a circle, it may be an oval or a square. A method of performing determination, a method of calculating the circularity of the entry content as the newly added content, instead of the circularity of the additional recording area, and the like can be applied.

  The analysis extraction unit 57 performs an analysis process on the entry form image 60 and extracts entry contents as newly added contents from each additional recording area. Further, the analysis extraction unit 57 performs an image extraction process on the additional recording area determined as the seal area, and extracts a seal image as newly added content. Each newly added content taken out is stored in the additional recording content storage unit 16 in association with the area identification number of the additional recording area.

  The display unit 58 includes a display, for example, and displays the contents of falsification in the entry form image 60 in which the tampered block 62 is detected based on the control of the control unit 59.

  The control unit 59 has a function of controlling each unit of the form processing apparatus 50.

Next, the operation of the form processing apparatus 50 according to the present embodiment will be described with reference to the drawings. Here, the flow of the form processing corresponding to the entry form image 60 shown in FIG. 9 will be described.
FIG. 10 is a flowchart (part 1) showing the operation of the form processing apparatus according to the second embodiment of the present invention, and FIG. 11 is a flowchart (part 2) showing the operation of the form processing apparatus according to the second embodiment of the present invention. ).

  When the form processing is started in the form processing apparatus 50, the image acquisition unit 11 first reads the form and acquires the entire image under the control of the control unit 59 (step S101). The acquired entire image is stored in the temporary storage unit 20.

  Subsequently, the control unit 59 controls the separation acquisition unit 12 to execute the separation process of the watermark image and the entry form image. The separation acquisition unit 12 reads the entire image from the temporary storage unit 20 and separates the watermark image and the form image based on the pixel density threshold (step S102). The separated watermark image and entry form image 60 (FIG. 9) are stored in the temporary storage unit 20.

  Next, the control unit 59 controls the analysis acquisition unit 52 to perform a watermark image analysis process. The analysis acquisition unit 52 reads a watermark image from the temporary storage unit 20 and performs a filtering process on the watermark image. Then, the analysis acquisition unit 52 acquires the feature information of the original form image 31 (FIG. 2A) included in the watermark image. The acquired feature information includes the division size of the original form image 31, the block identification data of each block divided from the original form image 31, the corresponding number of original pixels, and original shape data. 52 notifies the control unit 59 of the division size and stores the acquired block identification data, the corresponding number of original pixels, and original shape data in the temporary storage unit 20 (step S201).

  Upon receiving the notification of the division size, the control unit 59 controls the pixel number acquisition unit 13 to acquire the current pixel number from the entry form image 60 (FIG. 9). The number-of-pixels acquisition unit 13 reads the entry form image 60 from the temporary storage unit 20 and divides the entry form image 60 into blocks of the division size notified from the control unit 59, and also includes block identification data in these blocks. Is granted. Then, the pixel number acquisition unit 13 acquires the number of black pixels included in each block as the current pixel number (step S104). Each acquired number of current pixels is stored in the temporary storage unit 20 in association with the block identification data of the block.

  Subsequently, the control unit 59 controls the determination unit 53 to determine the presence / absence of entry contents in each block. Based on the pixel number difference and the pixel number difference threshold value, the determination unit 53 determines whether the block is a filled-in block (step S105).

  If the pixel number difference is equal to or greater than the pixel number difference threshold, the determination unit 53 determines that the block is a filled-in block having written content (step S105). Then, it is determined whether or not the number of original pixels corresponding to the block is 0 (step S106).

  When the number of original pixels is 0 (step S106), the determination unit 53 determines that the entry content included in the entry block is newly added content, and determines that the entry block is a block with additional writing. (Step S107). Then, the determination unit 53 notifies the control unit 59 of the block identification data of the block with additional writing together with the determination result.

  When receiving the block identification data from the determination unit 53, the control unit 59 stores the block identification data in the additional write content storage unit 16 (step S108).

  In step S106, when the number of original pixels is not 0, the determination unit 53 determines that the entry content included in the entry block is falsification content and determines that the entry block is a falsification block ( Step S202). Then, the determination unit 53 notifies the control unit 59 of the block identification data of the tampered block together with the determination result.

  When receiving the block identification data from the determination unit 53, the control unit 59 stores the block identification data in the falsification content storage unit 54 (step S203).

  In step S105, when the pixel number difference is less than the pixel number difference threshold, the determination unit 53 notifies the control unit 59 of the determination result. Receiving the notification, the control unit 59 controls the shape data acquisition unit 51 to acquire current shape data corresponding to the block (step S204).

  Subsequently, the control unit 59 notifies the determination unit 53 of the block identification data and the current shape data of the block, and determines whether or not the block has been tampered with. The determination unit 53 reads the corresponding original shape data from the temporary storage unit 20 based on the block identification data notified from the control unit 59. Then, the determination unit 53 calculates a shape difference based on the original shape data and the current shape data notified from the control unit 59, and compares the calculated shape difference with a stored shape difference threshold value ( Step S205).

  If the shape difference is greater than or equal to the shape difference threshold, the determination unit 53 determines that the block is a falsified block (step S202). Then, the determination unit 53 notifies the control unit 59 of the block identification data of the tampered block together with the determination result.

  When receiving the block identification data from the determination unit 53, the control unit 59 stores the block identification data in the falsification content storage unit 54 (step S203).

  In step S202, when the shape difference is less than the shape difference threshold, the determination unit 53 determines that the block is an unfilled block (step S206), and notifies the control unit 59 of the determination result.

  The control unit 59 notified of the determination result from the determination unit 53 determines whether there is an undetermined block (step S111). When there is an undetermined block, the control unit 59 causes the determination unit 53 to perform the processing from step S105 onward for the block.

  When it is determined that there is no undetermined block (step S111), the control unit 59 controls the region generation unit 55 to determine whether there is a block with additional writing (step S112). The area generation unit 55 determines the presence / absence of a block with additional writing based on the presence / absence of storage of block identification data in the additional writing content storage unit 16.

  When it is determined that there is a block with additional writing (step S112), the area generation unit 55 generates an adjacent block with additional writing as an additional writing area based on each block identification data stored in the additional writing content storage unit 16. Then, an area identification number is assigned to each additional recording area, and position data of each additional recording area is acquired (step S113). The area generation unit 55 generates additional writing areas 61-1 to 6-6 in the entry form image 60 (FIG. 9), assigns area identification numbers 1 to 6, and obtains position data of the additional writing areas 61-1 to 6-6. The area identification number and the corresponding position data are acquired and stored in the additional write content storage unit 16.

  Subsequently, the control unit 59 controls the seal stamp detection unit 56 to compare the region size of the additional recording region with the size threshold value (step S207). The seal stamp detection unit 56 first detects the region size “9” of the additional recording region 61-1, and determines that the size is less than the size threshold “16”.

  When the detected area size is less than the size threshold (step S207), the seal stamp detection unit 56 determines that the additional recording area 61-1 is not a seal stamp area, and notifies the control unit 59 of the determination result.

  Receiving the notification, the control unit 59 causes the analysis extraction unit 57 to extract the newly added content from the additional recording area 61-1. The analysis extraction unit 57 analyzes the additional writing area 61-1, extracts the newly added content “18”, and stores the newly added content in the additional writing content storage unit 16 in association with the area identification number “1” ( Step S116).

  Then, the control unit 59 causes the output unit 21 to read out and output the newly added content “18” and the corresponding position data from the additional write content storage unit 16 (step S117).

  Next, the control unit 59 determines whether or not there is a remaining additional recording area (step S118). If it is determined that there is a remaining additional recording area, the form processing apparatus 50 performs the processing from step S207 on the remaining additional recording area. Since the process flow for the write-once areas 61-2 and 3 is the same as the process flow for the write-once area 61-1, the description thereof will be omitted.

  Subsequently, the form processing apparatus 50 performs a process on the additional recording area 61-4. The seal stamp detection unit 56 detects the region size “32” for the additional recording region 61-4, and determines that the region size is equal to or larger than the size threshold “16” (step S207).

  When the area size of the additional recording area is equal to or larger than the size threshold (step S207), the seal detection unit 56 further calculates the circularity of the additional recording area, and compares the circularity with the circularity threshold (step S208). . Since the circularity “0.74” of the additional recording area 61-4 is equal to or greater than the circularity threshold “0.5”, the seal stamp detection unit 56 determines that the additional recording area 61-4 is a seal stamp area (Step S4). S209). The seal stamp detection unit 56 notifies the control unit 59 of the determination result.

  Receiving the notification, the control unit 59 causes the analysis extraction unit 57 to extract the seal image from the additional recording area 61-4. The analysis extraction unit 57 cuts out the seal image from the additional recording area 61-4, and stores the seal image in the additional recording content storage unit 16 in association with the area identification number “4” (step S210).

  And the control part 59 reads the position data corresponding to the seal image from the additional content storage part 16 by the output part 21, and outputs it (step S211).

  Next, the control unit 59 determines whether or not there is a remaining additional recording area (step S118), and causes the seal stamp detection unit 56 to detect the area size of the additional recording area 61-5 and compare it with the size threshold value (step S207). The seal stamp detection unit 56 detects the area size “24” of the additional recording area 61-5 and determines that the area size is equal to or larger than the size threshold (step S207).

  Furthermore, the seal stamp detection unit 56 calculates the circularity of the additional recording area 61-5, and compares the circularity with the circularity threshold (step S208). Since the circularity calculated from the additional recording area 61-5 is less than the circularity threshold, the seal detection unit 56 determines that the additional recording area 61-5 is not a seal area, and sends the determination result to the control unit 59. Notice.

  Upon receiving the notification, the control unit 59 causes the analysis extraction unit 57 to extract the newly added content “Okitaro” from the additional recording area 61-4, and associates the newly added content with the area identification number “5”. The information is stored in the additional write content storage unit 16 (step S116).

  And the control part 59 reads the position data corresponding to the newly added content "Taro Oki" from the additional record content storage part 16 by the output part 21, and outputs it (step S117).

  Subsequently, the control unit 59 determines whether or not there is a remaining additional recording area (step S118). If it is determined that there is a remaining additional recording area, the form processing apparatus 50 performs the processing from step S207 on the additional recording area 61-6. Since the process flow for the write-once area 61-6 is the same as the process flow for the write-once area 61-1, a description thereof will be omitted.

  When the output unit 21 outputs the newly added content “(3)” extracted from the additional recording area 61-6 from the additional recording content storage unit 16 and the corresponding position data (step S117), the control unit 59 further Then, the presence / absence of the remaining additional recording area is determined (step S118).

  If it is determined in step S118 that there is no remaining additional recording area, or if it is determined in step S112 that there is no additional recording area, the control unit 59 subsequently controls the area generation unit 55 to perform falsification. The presence / absence of a block is determined (step S212). The area generation unit 55 determines the presence / absence of a falsified block based on the presence / absence of block identification data in the falsification content storage unit 54.

  When it is determined that there is a falsified block (step S212), the area generation unit 55 generates an adjacent falsified block as a falsification area based on each block identification data stored in the falsification content storage unit 54. Then, an area identification number is assigned to each generated falsified area, and position data of each falsified area is acquired (step S213). The area generation unit 55 generates a falsified area from each falsified block 62 in the entry form image 60 (FIG. 9), assigns a region identification number 1, obtains position data of the falsified area, and identifies the area. The number and the corresponding position data are stored in the falsification content storage unit 54.

  Subsequently, the control unit 59 reads the entry form image 60 (FIG. 9) from the temporary storage unit 20 to warn that the falsification content has been detected, and first, the position corresponding to the region identification number “1”. The data is read from the falsification content storage unit 54 and the entry form image 60 including the falsification content is displayed on the display unit 58 (step S214). At that time, the display unit 58 performs highlighting such as changing the display color of the falsified area portion of the entry form image 60 based on the read position data.

  When the operator who viewed the display unit 58 performs predetermined processing on the falsified area and then resumes the form processing, the control unit 59 determines whether there is a remaining falsified area (step S215). If it is determined that there is a remaining falsified area, the form processing apparatus 50 performs the process of step S214 on the falsified area. If it is determined that there is no remaining falsified area, the form processing apparatus 50 ends the form processing.

  If it is determined in step S212 that there is no falsified block, the form processing by the form processing apparatus 50 ends.

  As described above, in the form processing apparatus according to the present embodiment, the presence / absence of entry contents is determined based on the shape difference in addition to the difference in the number of pixels, so that it is possible to detect format alteration with higher accuracy. When tampering is detected, an image of the tampered portion is displayed and processing by the operator is possible, so that for example, unauthorized application processing due to tampering with the application form is prevented. In addition, by calculating the circularity of the area detected as the additional recording area, it becomes possible to detect the seal area where the seal is imprinted. A form processing apparatus capable of preventing processing is realized.

FIG. 12 is a block diagram illustrating a configuration of a form processing apparatus according to the third embodiment of the present invention.
The form processing apparatus 70 according to the third embodiment is different from the first and second embodiments in that a ruled line search unit 71, a character search unit 72, and a setting unit 73 are added.
In the present embodiment, the same components as those in the first embodiment or the second embodiment are denoted by the same reference numerals, and detailed description thereof will be omitted.

  As illustrated in FIG. 12, the form processing apparatus 70 according to the present exemplary embodiment includes an image acquisition unit 11, a separation acquisition unit 12, a pixel number acquisition unit 13, a shape data acquisition unit 51, an analysis acquisition unit 52, a determination unit 53, and an additional note. A content storage unit 16, a falsification content storage unit 54, a region generation unit 55, a ruled line search unit 71, a character search unit 72, a setting unit 73, an analysis extraction unit 74, an output unit 75, a temporary storage unit 20, and a control unit 76 are provided. Yes.

The ruled line search unit 71 searches for a ruled line frame surrounding the additional write area based on the position data of each additional write area stored in the additional write content storage unit 16, and detects it as an area ruled line frame. Then, the ruled line search unit 71 notifies the control unit 76 of the area ruled line frame data such as the position of the detected area ruled line frame and the line width of each ruled line together with the area identification number of the additional recording area. The notified area ruled line frame data is stored in the additional content storage unit 16 in association with the notified area identification number.
Further, the ruled line search unit 71 detects a ruled line frame adjacent to the area ruled line frame in the entry form image as an adjacent ruled line frame, and sets adjacent ruled line frame data such as the position of the adjacent ruled line frame and the line width of each ruled line. The control unit 76 is notified together with the area identification number of the additional recording area. The notified adjacent ruled frame data is stored in the additional content storage unit 16 in association with the notified area identification number.
The ruled line search unit 71 detects the ruled line frame described above, for example, by pattern matching or the like.

FIG. 13 is a diagram (part 3) illustrating an example of an entry form image.
As shown in FIG. 13, the entry form image 80 includes additional recording areas 81-1 to 8-6 and a falsification area 82.

  For example, the ruled line search unit 71 searches the area ruled line frame 83 as a ruled line frame surrounding the additional write areas 81-1 to 8-3 in the form image 80 shown in FIG. Further, the ruled line search unit 71 detects an adjacent ruled line frame 84 as a ruled line frame adjacent to the area ruled line frame 83.

  Further, the ruled line search unit 71 searches for a ruled line frame surrounding the falsified area based on the position data of each falsified area stored in the falsified content storage unit 54 and detects it as a region ruled line frame. Then, the ruled line search unit 71 detects a ruled line frame adjacent to the area ruled line frame in the entry form image as an adjacent ruled line frame. The ruled line search unit 71 notifies the control unit 76 of the area ruled line frame data and the adjacent ruled line frame data together with the area identification number of the falsified area. The notified area ruled line frame data and adjacent ruled line frame data are stored in the falsification content storage unit 54 in association with the notified area identification number.

  For example, in the entry form image 80 (FIG. 13), the ruled line search unit 71 detects an area ruled line frame 85 and an adjacent ruled line frame 86 for the falsified area 82.

  The character search unit 72 sets an area surrounded by the ruled line frame as a search range based on each ruled line frame data stored in the additional record content storage unit 16 in the entry form image, and the characters included in the search range Explore data. First, the character search unit 72 searches for character data adjacent to the entry contents included in the additional writing area, that is, character data in the frame, using the area ruled line frame surrounding each additional writing area as a search range. In addition, the character search unit 72 searches for character data, that is, out-of-frame character data using the adjacent ruled line frame as a search range. The searched character data is sent to the control unit 76 together with the area identification number and stored in the additional content storage unit 16.

For example, in the entry form image 80 (FIG. 13), the character search unit 72 first adds, to the additional writing area 81-1, from the area ruled line frame 83, the in-frame character data “ Detect “Heisei” and “Year”. Further, the character search unit 72 detects out-of-frame character data “application date” from within the adjacent ruled line frame 84.
Similarly, in-frame character data “year” and “month” and out-of-frame character data “application date” are detected for the additional recording area 81-2, and the in-frame character data for the additional recording area 81-3. “Month” and “day” and out-of-frame character data “application date” are detected.
On the other hand, for the additional recording area 81-4, the character search unit 72 detects out-of-frame character data “stamping column” from within the adjacent ruled line frame 88. In-frame character data is not detected from within the area ruled line frame 87.

  Further, the character search unit 72 sets the inside of the adjacent ruled line frame as the search range based on the adjacent ruled line frame data stored in the falsification content storage unit 54 in the entry form image, and out-of-frame character data within the search range. Explore. The searched out-of-frame character data is sent to the control unit 76 together with the area identification number, and stored in the falsification content storage unit 54.

  For example, in the entry form image 80 (FIG. 13), the character search unit 72 applies the out-of-frame character data “application amount (please enter numbers 1 to 3) from the adjacent ruled line frame 86 to the falsified area 82. ”Is detected.

  The setting unit 73 sets, for each additional writing area, attribute information indicating an attribute of entry content included in the additional writing area. First, the setting unit 73 reads out in-frame character data and out-of-frame character data stored in correspondence with the region identification number from the additional writing content storage unit 16. When a plurality of in-frame character data is stored, the setting unit 73 selects the in-frame character data to be set as the attribute information based on a predetermined priority condition as the selected in-frame character data. Then, the setting unit 73 sets the attribute information “character data outside the frame: character data within the selection frame”. When either the character data inside the frame or the character data outside the frame is not stored, the setting unit 73 sets only the stored character data as attribute information. When neither the in-frame character data nor the out-of-frame character data is stored, the setting unit 73 sets the attribute information “not set”. The set attribute information is stored in the additional content storage unit 16 in association with the area identification number.

  In this embodiment, it is assumed that a priority condition is set in the setting unit 73 so as to preferentially select character data of “year”, “month”, and “day”. In addition, it is assumed that high priority conditions are set in order for each character data “year”, “month”, and “day”. In other words, priority conditions are set so that “month” rather than “year” and “day” than “month” are selected as character data in the selection frame.

  For example, in the entry form image 80 (FIG. 13), the setting unit 73 gives priority to the additional writing area 81-1 from the in-frame character data “Heisei” and “Year” stored in the additional writing content storage unit 16. Based on the condition, “year” is selected as the character data in the selection frame. Then, the setting unit 73 sets the attribute information “application date: year” for the additional recording area 81-1. Similarly, the setting unit 73 sets attribute information “application date: month” for the additional recording area 81-2 and attribute information “application date: day” for the additional recording area 81-3. The setting unit 73 sets the attribute information “stamping column” for the additional recording area 81-4.

  The setting unit 73 similarly sets attribute information for the falsified area. The setting unit 73 reads out-of-frame character data stored corresponding to the region identification number from the falsification content storage unit 54 and sets the attribute information “out-of-frame character data”.

  For example, in the entry form image 80 (FIG. 13), the setting unit 73 sets the attribute information “application amount (please enter numbers 1 to 3)” for the falsified area 82.

  The analysis extraction unit 74 performs an analysis process on the entry form image, and extracts entry contents as newly added contents from each additional recording area. The analysis extraction unit 74 refers to the attribute information stored in the additional writing content storage unit 16 and stamps the additional writing area in which the attribute information such as “stamping column” and “sealing stamp column” is set. It is determined that the area is a stamp area, and an image cutting process is performed on the additional area to extract a seal image as newly added content. For other additional recording areas, OCR processing is performed to extract newly added contents. The extracted newly added content is stored in the additional write content storage unit 16 in association with the area identification number.

  In addition, the analysis extraction unit 74 extracts entry contents as falsification contents from the falsification area. Based on the area ruled line frame data stored in the falsification content storage unit 54, the analysis extraction unit 74 performs image cutout processing on the area ruled line frame, and extracts an image including the falsification contents as a falsified image. The extracted altered image is stored in the altered content storage unit 54 in association with the area identification number.

  The output unit 75 assigns a form identification number to each entry form image, and outputs the entry contents read from the entry form image and the corresponding attribute information together with the form identification number. The output unit 75 reads the newly added content and the corresponding attribute information from the additional write content storage unit 16 and outputs it as output data. Further, the output unit 75 reads the falsification content and the corresponding attribute information from the falsification content storage unit 54, and outputs it as output data.

FIG. 14 is a diagram illustrating an example of output data.
FIG. 14A shows an output example of newly added contents and corresponding attribute information, and FIG. 14B shows an output example of a falsified image and corresponding attribute information.

Next, the operation of the form processing apparatus 70 of this embodiment will be described with reference to the drawings. Here, the flow of the form processing corresponding to the entry form image 80 shown in FIG. 13 will be described.
FIG. 15 is a flowchart (part 1) showing the operation of the form processing apparatus according to the third embodiment of the present invention, and FIG. 16 is a flowchart (part 2) showing the operation of the form processing apparatus according to the third embodiment of the present invention. ).

  When the form processing is started in the form processing apparatus 70, the image acquisition unit 11 reads the form and acquires the entire image (step S101), and then adds each block to the block with additional writing (step S107) and the block with tampering (step S202), or a block without entry (step S206). Since the flow of processing until it is determined that there is no undetermined block (step S111) is the same as that in the second embodiment (FIG. 10), description thereof is omitted.

  If it is determined in step S111 (FIG. 15) that there is no undecided block, the control unit 76 controls the region generation unit 55 to determine whether there is a block with additional writing (step S112). The area generation unit 55 determines the presence / absence of a block with additional writing based on the presence / absence of storage of block identification data in the additional writing content storage unit 16.

  When it is determined that there is a block with additional writing (step S112), the area generation unit 55 generates an adjacent block with additional writing as an additional writing area based on each block identification data stored in the additional writing content storage unit 16. Then, an area identification number is assigned to each generated additional recording area, and position data of each additional recording area is acquired (step S113). The area generation unit 55 generates additional write areas 81-1 to 8-6 in the entry form image 80 (FIG. 13), assigns area identification numbers 1 to 6, and obtains position data of the additional write areas 81-1 to 8-6. The area identification number and the corresponding position data are acquired and stored in the additional write content storage unit 16.

  Subsequently, the control unit 76 controls the ruled line search unit 71 to search for the area ruled line frame and the adjacent ruled line frame of the additional writing area (step S301). The ruled line search unit 71 first reads the position data of the additional writing area 81-1 stored in the additional writing content storage unit 16 in order to search for the area ruled line frame and the adjacent ruled line frame corresponding to the additional writing area 81-1. Based on the position data, the ruled line search unit 71 searches the area ruled line frame 83 and the adjacent ruled line frame 84, and sends the area ruled line frame data and the adjacent ruled line frame data to the control unit 76 together with the area identification number “1”. Notice. The control unit 76 stores the notified area ruled line frame data and adjacent ruled line frame data in the additional content storage unit 16 in association with the area identification number “1”.

  Next, the control unit 76 controls the character search unit 72 to search for the character data within the frame and the character data outside the frame in the additional writing area (step S302). The character search unit 72 searches for the character data in the frame in the area ruled line frame 83 corresponding to the additional recording area 81-1, and detects the character data “Heisei” and “year” in the vicinity of the additional recording area 81-1. To do. Further, the character search unit 72 searches for the out-of-frame character data in the adjacent ruled line frame 84 and detects the out-of-frame character data “application date”. Then, the character search unit 72 notifies the control unit 76 of the detected character data “Heisei” and “year” and the out-of-frame character data “application date” together with the area identification number “1”. The control unit 76 stores the notified in-frame character data and out-of-frame character data in the additional content storage unit 16 in association with the region identification number.

  Subsequently, the control unit 76 controls the setting unit 73 to set attribute information for the additional recording area (step S303). The setting unit 73 reads the frame character data “Heisei” and “year” corresponding to the area identification number “1” from the additional record content storage unit 16 and selects the character data “year” in the selection frame based on a predetermined priority condition. ”Is selected. Further, the setting unit 73 reads out-of-frame character data “application date” corresponding to the area identification number “1” from the additional writing content storage unit 16. Then, the setting unit 73 sets “application date: year” as attribute information indicating the attribute of the content to be entered in the additional writing area 81-1, and stores it in the additional writing content storage unit 16 in association with the area identification number “1”. Remember.

  Next, the control unit 76 controls the analysis extracting unit 74 to execute extraction of newly added contents from the additional recording area. The analysis extraction unit 74 refers to the attribute information “application date: year” stored in the additional record content storage unit 16 corresponding to the region identification number “1”, and then adds the newly added content “18” from the additional record region 81-1. ”Is extracted, and the newly added content is associated with the area identification number“ 1 ”and stored in the additional content storage unit 16 (step S304).

  Subsequently, the control unit 76 determines whether or not there is a remaining additional recording area (step S118). If it is determined that there is a remaining additional recording area, the form processing apparatus 70 performs the processing from step S301 on the remaining additional recording area. Since the process flow for the write-once areas 81-2 and 3 is the same as the process flow for the write-once area 81-1, description thereof will be omitted.

  Subsequently, the form processing apparatus 70 performs the processing after step S301 on the additional recording area 81-4. That is, the ruled line search unit 71 searches the area ruled line frame 87 and the adjacent ruled line frame 88 corresponding to the additional recording area 81-4, and the area ruled line frame data and the adjacent ruled line frame data are additionally recorded corresponding to the area identification number “4”. It is stored in the content storage unit 16 (step S301).

  Next, the character search unit 72 searches the area ruled line frame 87 and the adjacent ruled line frame 88 for character data, and detects the out-of-frame character data “marking field” from the adjacent ruled line frame 88 (step S302). Then, the setting unit 73 sets the attribute information “stamping column” of the additional recording area 81-4 (step S303).

  Subsequently, the control unit 76 causes the analysis extraction unit 74 to extract the newly added content from the additional recording area 81-4. The analysis extraction unit 74 refers to the attribute information of the additional writing area 81-4 stored in the additional writing content storage unit 16, and determines that the additional writing area 81-4 in which the attribute information “stamping column” is set is a seal area. Then, the image cut-out process is performed on the additional write area 81-4, the seal image is taken out as the newly added content, and is stored in the additional write content storage unit 16 in association with the area identification number “4” (step S304). .

  Subsequently, the control unit 76 determines whether or not there is a remaining additional recording area (step S118). If it is determined that there is a remaining additional recording area, the form processing apparatus 70 adds the remaining additional recording areas 81-5, 6 to On the other hand, the process after step S301 is implemented.

  If it is determined in step S118 that there is no remaining additional recording area, the control unit 76 causes the output unit 75 to read the newly added content and the corresponding attribute information from the additional recording content storage unit 16, and FIG. As shown in FIG. 5, the newly added contents and the attribute information are output in association with the form identification number of the form (step S305). The output form identification number, newly added contents, and attribute information are stored in a database (not shown).

  When the newly added content is output in step S305, or when it is determined in step S112 that there is no block with additional writing, the control unit 76 controls the region generation unit 55 to determine whether there is a block with falsification. (Step S212). The area generation unit 55 determines the presence / absence of a falsified block based on the presence / absence of storage of block identification data in the falsification content storage unit 54.

  When it is determined that there is a falsified block (step S212), the area generation unit 55 generates an adjacent falsified block as a falsification area based on each block identification data stored in the falsification content storage unit 54. Then, an area identification number is assigned to each generated falsified area, and position data of each falsified area is acquired (step S213). The area generation unit 55 generates a falsified area 82 in the entry form image 80 (FIG. 13), assigns the area identification number 1, acquires position data of the falsified area 82, and obtains the area identification number and the corresponding position. The data is stored in the falsification content storage unit 54.

  Subsequently, the control unit 76 controls the ruled line search unit 71 to search for an area ruled line and an adjacent ruled line frame corresponding to the falsified area (step S306). The ruled line search unit 71 searches the area ruled line frame 85 and the adjacent ruled line frame 86 based on the position data of the falsified area 82 stored in the falsification content storage unit 54, and the area ruled line frame data and the adjacent ruled line frame data are The control unit 76 is notified together with the area identification number “1”. The control unit 76 stores the area ruled line frame data and the adjacent ruled line frame data in the falsification content storage unit 54 in association with the area identification number.

  Next, the control unit 76 controls the character search unit 72 to search for out-of-frame character data corresponding to the falsified area (step S307). The character search unit 72 searches for the out-of-frame character data in the adjacent ruled line frame 86 corresponding to the falsified area 82, and detects the detected out-of-frame character data “application amount (please enter the numbers 1 to 3)”. The control unit 76 is notified together with the area identification number “1”. The control unit 76 stores the notified out-of-frame character data in the falsification content storage unit 54 in association with the area identification number.

  Subsequently, the control unit 76 controls the setting unit 73 to set attribute information for the falsified area (step S308). The setting unit 73 reads out-of-frame character data corresponding to the falsification area 82 from the falsification content storage unit 54 and sets the attribute information “application amount (please enter the numbers 1 to 3)”. The set attribute information is stored in the falsification content storage unit 54 in association with the area identification number “1”.

  Next, the control unit 76 controls the analysis extraction unit 74 to execute extraction of falsification contents from the falsification area. The analysis extraction unit 74 reads out the area ruled line frame data corresponding to the falsified area 82 from the falsification content storage unit 54, performs image cutting processing on the area ruled line frame 85, and extracts a falsified image including the falsification contents. . Then, the analysis extraction unit 74 stores the extracted falsified image in the falsification content storage unit 54 in association with the region identification number “1” (step S309).

  Subsequently, the control unit 76 determines whether or not there is a remaining falsified area (step S215). If it is determined that there is a remaining falsified area, the form processing apparatus 70 performs the processing from step S306 on the remaining falsified area.

  When it is determined that there is no remaining falsified area (step S215), the control unit 76 causes the output unit 75 to read the falsified image including the falsified content and the corresponding attribute information from the falsified content storage unit 54, and FIG. As shown in (b), the falsified image and attribute information are output in association with the form identification number of the form (step S310). The output form identification number, falsified image, and attribute information are stored in a database (not shown). Thereby, the form processing apparatus 70 ends the form processing.

  If it is determined in step S212 that there is no falsified block, the form processing by the form processing apparatus 70 ends.

  As described above, the form processing apparatus according to the present embodiment detects attribute information indicating the attribute of the entry contents and outputs the attribute information in association with the entry contents, so that a plurality of entry contents are automatically read from a large number of forms. Thereafter, the entry content can be converted into data in a form that can be aggregated for each attribute information and output, and the applicable range is expanded. In addition, since the falsification image including the falsification contents and the attribute information of the falsification contents are output, it is possible to generate a large number of forms without detecting the falsification contents and carrying out human processing. After processing, it is possible to refer to the output result and perform processing for the form including the contents of falsification again, thereby realizing efficient form processing.

  In this embodiment, the attribute information is set by reading from the vicinity of the entry contents in the form, but is not limited to this. For example, it is possible to adopt a configuration in which an image of a form is displayed for a form for which attribute information cannot be set, and attribute information is selected or input.

It is a block diagram which shows the structure of the form processing apparatus which concerns on Example 1 of this invention. It is a figure which shows an example of a form. It is a figure explaining the separation process of an entry form image and a watermark image. It is explanatory drawing of the present pixel number acquisition process in an entry form image. It is FIG. (1) which shows an example of an entry form image. It is a flowchart (the 1) which shows operation | movement in Example 1 of the form processing apparatus which concerns on this invention. It is a flowchart (the 2) which shows operation | movement in Example 1 of the form processing apparatus which concerns on this invention. It is a block diagram which shows the structure of the form processing apparatus which concerns on Example 2 of this invention. It is FIG. (2) which shows an example of an entry form image. It is a flowchart (the 1) which shows operation | movement in Example 2 of the form processing apparatus which concerns on this invention. It is a flowchart (the 2) which shows operation | movement in Example 2 of the form processing apparatus which concerns on this invention. It is a block diagram which shows the structure of the form processing apparatus which concerns on Example 3 of this invention. It is a figure (the 3) which shows an example of an entry form image. It is a figure which shows an example of output data. It is a flowchart (the 1) which shows operation | movement in Example 3 of the form processing apparatus which concerns on this invention. It is a flowchart (the 2) which shows operation | movement in Example 3 of the form processing apparatus which concerns on this invention.

Explanation of symbols

10, 50, 70 Form processing device 11 Image acquisition unit 12 Separation acquisition unit 13 Pixel number acquisition unit 14, 52 Analysis acquisition unit 15, 53 Determination unit 19, 57, 74 Analysis extraction unit 21, 75 Output unit 51 Shape data acquisition unit 72 Character search part

Claims (9)

A form that is printed by dividing the original form image into a plurality of blocks, obtaining each feature data corresponding to the position of each block , and embedding the watermark image generated based on the feature data in the original form image A form processing device for reading entry contents from
An image acquisition unit that reads the form that has been filled in to obtain an overall image;
A separation acquisition unit that separates and acquires the watermark image and the entry form image including the entry content and the original form image from the acquired whole image;
Analyzing the acquired watermark image, and obtaining a plurality of the feature data as original feature data respectively;
A feature acquisition unit that acquires each feature data corresponding to the position of each block from the acquired form image as current feature data;
E Bei a determination unit for determining presence or absence of the entry contents by comparing each original feature data and said each current feature data
The density of each pixel constituting the watermark image is based on a predetermined density threshold, and has a gradation density different from the density of each pixel constituting the original form image and the entry content,
The separation acquisition unit acquires density data indicating the density of each pixel constituting the entire image, and separates the watermark image and the entry form image based on the predetermined density threshold. Form processing device to do. The plurality of original feature data, the number of black pixels included in the plurality of blocks as each position obtained by dividing the original form image into a grid of a predetermined size is indicated as the number of original pixels
  The plurality of current feature data acquired by the feature acquisition unit includes the number of black pixels included in the plurality of blocks as the positions obtained by dividing the entry form image into a grid of the predetermined size. As the current number of pixels,
  When the difference data indicating the difference between the number of original pixels and the number of current pixels in the same block is equal to or greater than a predetermined difference threshold, the determination unit determines that the block is a filled block having the filled content Do
  The form processing apparatus according to claim 1. When it is determined that the block is filled in, the determination unit further determines the number of original pixels corresponding to the filled block. If the number of original pixels is 0, the determination unit includes the filled block. 3. The form according to claim 2, wherein it is determined that the entered content is newly added content, and if the original pixel number is not 0, it is determined that the entered content included in the entered block is falsified content. Processing equipment. The form processing apparatus according to claim 2, further comprising an analysis extraction unit that analyzes the entry form image and extracts the entry contents from each entry block. In the entry form image, a character search unit for searching character data from a block adjacent to the entry block,
  An output unit for outputting the searched character data as attribute information indicating the attribute of the entry content in association with the entry content;
  The form processing apparatus according to claim 4. A form that is printed by dividing an original form image into a plurality of blocks, obtaining each feature data corresponding to the position of each block, and embedding a watermark image generated based on the feature data in the original form image A form processing device for reading entry contents from
  An image acquisition unit that reads the form that has been filled in to obtain an overall image;
  A separation acquisition unit that separates and acquires the watermark image and the entry form image including the entry content and the original form image from the acquired whole image;
  Analyzing the acquired watermark image, and obtaining a plurality of the feature data as original feature data respectively;
  A feature acquisition unit that acquires each feature data corresponding to the position of each block from the acquired form image as current feature data;
  A determination unit that determines the presence or absence of the entry content by comparing the original feature data and the current feature data.
A plurality of the original feature data uses a shape extraction filter for extracting a specific shape for a plurality of blocks as the positions obtained by dividing the original form image into a grid of a predetermined size. The original shape data calculated by
  The plurality of the current feature data acquired by the feature acquisition unit is the shape extraction for the plurality of blocks as the respective positions obtained by dividing the form image into a grid of the predetermined size. Consists of current shape data calculated using a filter,
  When the difference data indicating the difference between the original shape data and the current shape data in the same block is equal to or greater than a predetermined difference threshold, the determination unit determines that the block is a filled block having the filled content. Do
  A form processing apparatus characterized by that. The form processing apparatus according to claim 6, wherein the shape extraction filter is a Gabor filter. The form processing apparatus according to claim 6, wherein if it is determined that the entry block is present, the determination unit further determines that the entry content is falsification content. The form processing apparatus according to claim 6, further comprising: an analysis extraction unit that analyzes the entry form image and extracts the entry contents from each entry block.

Images