JP2013255110A - Image processing device, method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the accuracy of determination on a correct combination of partial images even when a plurality of partial images obtained by reading divided sections of a document page do not have a distinguished difference in edges of the image.SOLUTION: A plurality of partial images obtained by reading divided sections of a document page by a scanner are stored in a scan image storage unit 12. A verification information acquisition unit 14 acquires verification information which helps to determine a combination of partial images which, among combinations of partial images, does not hold as a combination of those on the original document page. An exclusion combination determination unit 18 determines a combination of partial images which does not hold on the basis of the verification information. A likelihood arithmetic unit 20 calculates a likelihood of adjacency on the original document page with regard to a plurality of partial image combinations. A combination determination unit 24 determines, from the plurality of partial image combinations excluding those which were determined not to hold by the exclusion combination determination unit 18, a combination of partial images adjacent on the original document page on the basis of the calculated likelihood.

Description

  The disclosed technology relates to an image processing apparatus, an image processing method, and an image processing program.

  When reading an original with a scanner, if the original size that can be read by the scanner is smaller than the original size to be read, the following reading methods are known. For example, when the original size that can be read by the scanner is A4 and the original size to be read is A3, the A2 size original is folded in half and sandwiched at a predetermined position on a dedicated sheet. This is a method of scanning and connecting scanned images. In the second reading method, for example, when the original size A4 is readable by the scanner and the original size to be read is A3, the scanner is set to the non-separating paper feeding mode, and the A3 size original folded in half is read. This is a method of performing scanning on both sides and connecting scanned images.

  However, in both the first reading method and the second reading method, after the document is folded in half, it is necessary to manually put the documents into the scanner one by one. Therefore, even if there are a large number of documents to be read, the ADF (Auto Document Feeder) cannot be used, which requires a lot of work. Therefore, as a third reading method, the original is cut into a size that can be read by the scanner (for example, A3 size is half A4 size), scanned using ADF, and then the read image is restored to the original size image. A method of configuring (connecting) is conceivable. However, in the third reading method, in order to reconstruct the read image into the original size image, it is necessary to perform a process of determining a correct combination of the read images.

  In relation to the above, a technique has been proposed in which images that are divided and read are automatically joined together. This technology searches each document data divided and stored in the storage unit, detects a characteristic line segment, symbol, or data change amount existing at the end of the image, and the characteristic data amount substantially matches. Recognizes the joint part from the side.

Japanese Patent Laid-Open No. 7-23204

  However, the above technique is based on the premise that features such as line segments and symbols exist near the image edge of each image, and that there are clear differences in the features existing near the image edge of each image. It is said. For this reason, the above-described technique is used when there is no clear difference in the characteristics existing near the image edge of each image, such as an image obtained by dividing and reading a document with a sentence or a table on the entire surface. Another problem is that the correct combination of adjacent images on the original document cannot be determined.

  As one aspect, the disclosed technique is a method for correctly combining partial images even when there is no clear difference between image edges of a plurality of partial images obtained by dividing a document page into a plurality of portions. The purpose is to improve the judgment accuracy.

  The disclosed technology obtains verification information for determining that a combination is not formed as one page of a document from a partial image that is one page of a document when combined with another partial image. Department. In addition, the disclosed technique includes a determination unit that determines a combination that is not established as a combination on the original page of the document based on the verification information acquired by the acquisition unit. The disclosed technique includes a combination determination unit that excludes combinations determined to be not established by the determination unit among arbitrary combinations of the plurality of partial images.

  As one aspect, the disclosed technique is a method for correctly combining partial images even when there is no clear difference between image edges of a plurality of partial images obtained by dividing a document page into a plurality of portions. The determination accuracy can be improved.

It is a functional block diagram which shows an example of the image processing apparatus demonstrated by embodiment. And FIG. 20 is a schematic block diagram illustrating an example of a computer that functions as an image processing apparatus. (A) is a schematic configuration diagram showing an example of a document reading unit of the scanner, and (B) is a table showing an example of front / back identification information recorded by the scanner. FIG. 6 is an image diagram illustrating an example of reading an original by dividing (cutting) the original. It is a flowchart which shows the image combination process demonstrated in 1st Embodiment. It is a graph which shows an example of partial image identification information. It is a chart which shows an example of the initial state of a likelihood calculation chart. It is a graph which shows an example of the likelihood calculation chart after excluding the combination which is not materialized as a combination on the original page of a manuscript. It is an image figure which shows an example of the block of the power spectrum calculation object which exists inside a side. It is a chart which shows an example of the likelihood calculation chart which passed through the calculation of likelihood. It is a graph which shows an example of the determination result of the group of the edge | side of a connection object. It is a flowchart which shows the image combination process which concerns on 2nd Embodiment. In 2nd Embodiment, it is a chart which shows an example of the likelihood calculation chart which passed through the calculation of likelihood. It is a flowchart which shows the combination verification process which is not materialized based on 2nd Embodiment. It is a graph which shows an example of the top and bottom direction information for every partial image in 3rd Embodiment. It is a flowchart which shows the combination verification process which is not materialized concerning 3rd Embodiment. It is a flowchart which shows the verification information acquisition process which concerns on 4th Embodiment. It is an image figure which shows an example of the page number area | region designation | designated screen in 4th Embodiment. It is a table | surface which shows an example of the judgment result of the presence or absence of the page number for every partial image in 4th Embodiment. It is a flowchart which shows the combination verification process which is not materialized based on 4th Embodiment. It is a flowchart which shows the verification information acquisition process which concerns on 5th Embodiment. It is an image figure which shows an example of the page number area | region designation | designated screen in 5th Embodiment. It is a chart which shows an example of page number field position information in a 5th embodiment. It is a flowchart which shows the combination verification process which is not materialized based on 5th Embodiment.

  Hereinafter, an example of an embodiment of the disclosed technology will be described in detail with reference to the drawings.

[First Embodiment]
FIG. 1 shows an image processing apparatus 10 according to the present embodiment. The image processing apparatus 10 cuts a document page into a size that can be read by the scanner (for example, A3 size is a half A4 size) and reads the original page, and then connects the obtained partial image to the original size image. In this case, the apparatus performs processing for determining a correct combination of partial images. The image processing apparatus 10 includes a scanned image storage unit 12, a verification information acquisition unit 14, a verification information storage unit 16, an exclusion combination determination unit 18, a likelihood calculation unit 20, a likelihood storage unit 22, a combination determination unit 24, and a combination storage unit. 26.

  The scan image storage unit 12 acquires image data of a plurality of partial images obtained by cutting each page of the document into a size that can be read by the scanner and reading the scanned image from the scanner, and the acquired plurality of partial images. Is stored. The verification information acquisition unit 14 is capable of determining a combination that is not established as a combination of partial images on the original page of the original document from among arbitrary combinations of a plurality of partial images whose image data is stored in the scanned image storage unit 12. Get information. A specific example of the verification information will be described later. The verification information storage unit 16 stores the verification information acquired by the verification information acquisition unit 14.

  Based on the verification information acquired by the verification information acquisition unit 14 and stored in the verification information storage unit 16, the exclusion combination determination unit 18 is a part on the original page of the original document among arbitrary combinations of a plurality of partial images. A combination that is not established as a combination of images is determined. Based on the image data of the plurality of partial images stored in the scan image storage unit 12, the likelihood calculation unit 20 represents the likelihood that each combination of the plurality of partial images is adjacent on the original page of the document. Calculate the likelihood. The likelihood storage unit 22 stores the likelihood calculated by the likelihood calculation unit 20 for each combination of a plurality of partial images.

  Based on the combination that is determined as the combination of the partial images of the original page of the original determined by the exclusion combination determining unit 18 and the likelihood calculated by the likelihood calculating unit 20, the combination determining unit 24 Determine the correct combination of adjacent partial images on the page. The combination storage unit 26 stores the correct combination of the partial images adjacent to each other on the original page of the original determined by the combination determination unit 24.

  The verification information acquisition unit 14 is an example of an acquisition unit in the disclosed technology, and the exclusion combination determination unit 18 is an example of a determination unit in the disclosed technology. The likelihood calculating unit 20 is an example of a likelihood calculating unit in the disclosed technique, and the combination determining unit 24 is an example of a combination determining unit in the disclosed technique.

  The image processing apparatus 10 can be realized, for example, by a computer 30 shown in FIG. The computer 30 includes a CPU 32, a memory 34, a nonvolatile storage unit 36, a display 38, an input device 40 such as a keyboard and a mouse, and an external I / F unit 42. The CPU 32, the memory 34, the storage unit 36, the display 38, the input device 40, and the external I / F unit 42 are connected to each other via a bus 45. A scanner 44 is connected to the external I / F unit 42.

  The storage unit 36 can be realized by an HDD (Hard Disk Drive), a flash memory, or the like. The storage unit 36 as a recording medium stores an image combination program 46 for causing the computer 30 to function as the image processing apparatus 10, and includes a scan image storage area 56, a verification information storage area 58, a likelihood storage area 60, and a combination storage area. 62 is provided. The CPU 32 reads out the image combination program 46 from the storage unit 36 and develops it in the memory 34, and sequentially executes processes included in the image combination program 46.

  The image combination program 46 includes a verification information acquisition process 48, an exclusion combination determination process 50, a likelihood calculation process 52, and a combination determination process 54. The CPU 32 operates as the verification information acquisition unit 14 illustrated in FIG. 1 by executing the verification information acquisition process 48. The CPU 32 operates as the exclusion combination determination unit 18 illustrated in FIG. 1 by executing the exclusion combination determination process 50. The CPU 32 operates as the likelihood calculating unit 20 shown in FIG. 1 by executing the likelihood calculating process 52. The CPU 32 operates as the combination determination unit 24 illustrated in FIG. 1 by executing the combination determination process 54.

  When the image processing apparatus 10 is realized by the computer 30, the scan image storage area 56 functions as the scan image storage unit 12 illustrated in FIG. 1, and the verification information storage area 58 functions as the verification information storage unit 16 illustrated in FIG. . The likelihood storage area 60 functions as the likelihood storage section 22 shown in FIG. 1, and the combination storage area 62 functions as the combination storage section 26 shown in FIG. As a result, the computer 30 that has executed the image combination program 46 functions as the image processing apparatus 10. The image combination program 46 is an example of an image processing program in the disclosed technology.

  Note that the image processing apparatus 10 can be realized by, for example, a semiconductor integrated circuit, more specifically, an ASIC (Application Specific Integrated Circuit) or the like.

  Further, the scanner 44 according to the first embodiment reads a pair of pages obtained by reading the front and back surfaces of the same partial document when the front and back surfaces of the partial document obtained by dividing the page of the document into a plurality of parts are read. Front / back identification information for identifying a partial image is recorded. For example, as shown in FIG. 3A, the scanner 44 has a configuration in which the reading unit 66 conveys a document by a pair of conveyance rollers 68 and 70 and reading sensors 72 and 74 are provided on both sides of the document conveyance path. The front and back identification information is added to the partial images simultaneously read by the reading sensors 72 and 74. For example, as shown in FIG. 3B, the front / back identification information is information indicating that the image data Image0 and Image1, and Image2 and Image3 of the partial image are in the front / back positional relationship in the original document.

  The front / back identification information can be recorded by a scanner that is not configured to simultaneously read both the front and back sides of the document. For example, even a scanner equipped with a single reading sensor and a reversing mechanism for reversing the front and back of a document can recognize that the reversing mechanism has been activated to reverse the front and back of the document. Even a scanner configured to read can record front / back identification information. Furthermore, even if the scanner is configured to place a document to be scanned by the user on the document placement surface, “Please set the front side”, “Set the back side”, etc. By sequentially displaying these messages, the front and back of the document can be recognized. Therefore, it is possible to record the front / back identification information even with the scanner having the above-described configuration.

  Next, the operation of the first embodiment will be described. When the page size of the document is larger than the size that can be read by the scanner 44, the page of the document is cut into a plurality of partial documents having a size smaller than the size that can be read by the scanner 44 and read by the scanner 44. In the following, the document size is A3 and the readable size of the scanner 44 is A4. As shown in FIG. 4, the A3 document printed on both sides is cut into two partial documents and printed on one side. A case where the A3 original is cut into three partial originals and each of them is read will be described as an example.

  In the example of FIG. 4, an A3 size document printed on both sides is cut into two A4 size partial documents to obtain four A4 size partial images, and A3 size printed on one side. The size original is cut into three partial originals to obtain three partial images. Hereinafter, individual partial images are referred to as Image0 to Image6 to distinguish them. In the example of FIG. 4, since the document is cut in the vertical direction (direction parallel to the short side of the original document), the sides that are candidates for the boundary connecting the partial images are the left and right sides of the partial image in FIG. The left and right sides of each partial image are referred to as h0 to h13, respectively.

  Note that when reading a partial original obtained by cutting an A3 size original printed on one side into three, both sides are actually read, but the blank page determination process performed by the scanner 44 causes the blank page on the back side to be read. Since the image is deleted, it is not shown in FIG. Although FIG. 4 shows an example in which the document is cut only in the vertical direction, the present invention is also applied to a case where the document is cut in the vertical direction and the horizontal direction and the sides that are candidates for the boundary connecting the partial images are four sides. Is possible.

  When the reading of the partial original by the scanner 44 is completed, the image processing apparatus 10 performs an image combination process shown in FIG. In step 100 of the image combination process, the scan image storage unit 12 acquires the image data of the partial image obtained by reading the partial original by the scanner 44 from the scanner 44 and stores the acquired image data of the partial image. Further, the scan image storage unit 12 gives partial image identification information to each acquired partial image. As shown in FIG. 6 as an example, the partial image identification information given to the partial image includes an identifier for identifying each partial image, an attribute of image data, and an identifier of a side that is a candidate for a boundary connecting the partial images. (H0 to h13 in the example of FIG. 6) is included.

  In the next step 102, the verification information acquisition unit 14 acquires verification information that can determine a combination of partial images that is not established as a combination on the original page of the document. In the first embodiment, it is possible to determine that a pair of partial images obtained by reading the front and back sides of the same partial document are not adjacent images on the original page of the document. As shown in FIG. 3B, front / back identification information as shown in FIG. The front / back identification information shown in FIG. 3B corresponds to the example of FIG. The verification information acquired by the verification information acquisition unit 14 is stored in the verification information storage unit 16.

  In step 104, the exclusion combination determination unit 18 is not adjacent on the original page of the document based on the verification information (front / back identification information) acquired by the verification information acquisition unit 14 and stored in the verification information storage unit 16. A combination in which a side of a partial image is not established is determined.

  For example, according to the front / back identification information shown in FIG. 3B, the partial image Image0 and the partial image Image1 are partial images obtained by reading the front and back of the same partial document, and have a front-back relationship. Therefore, it can be determined that the sides h0 and h1 of the partial image Image0 and the sides h2 and h3 of the partial image Image1 are not adjacent to each other on the original page of the document. Also, according to the front / back identification information shown in FIG. 3B, the partial image Image2 and the partial image Image3 are also partial images obtained by reading the front and back of the same partial document, and have a front-back relationship. Accordingly, it can be determined that the sides h4 and h5 of the partial image Image2 and the sides h6 and h7 of the partial image Image3 are not adjacent to each other on the original page of the document.

  When the exclusion combination determination unit 18 determines a combination of sides of partial images that are not established as a combination of the original pages of the document, the exclusion combination determination unit 18 performs processing for excluding the determined combination that is not established from the target of the likelihood calculation by the likelihood calculation unit 20. Do. FIG. 7 shows an initial state of a likelihood calculation chart in which the likelihood calculation unit 20 displays a list of combinations of edges on which the likelihood is calculated. The exclusion combination determination unit 18 is configured to exclude the determined combination from the combinations of the edges shown in the likelihood calculation chart shown in FIG. 7 as processing for excluding the determined combination from the object of the likelihood calculation by the likelihood calculation unit 20. Information indicating that the combination is not determined is set for each combination of sides determined not to be calculated. FIG. 8 shows a cell corresponding to a combination of sides in which information indicating that the information is not a calculation target is set with a symbol “x”. As described above, prior to the likelihood calculation by the likelihood calculation unit 20, the combination of the sides of the partial images that are not established as the combination on the original page of the document is excluded from the calculation target. Calculation load is reduced. In addition, since it is possible to exclude the combination determination unit 24 from erroneously determining the combination that is not established, the accuracy of calculating the correct combination is improved.

  In the next step 106, the likelihood calculating unit 20 sets a boundary between the pair of partial images on the original page of the original for a combination of sides that are not set to be calculated among the combinations of the sides of the partial images. Likelihoods that represent the likelihood of being in a positional relationship adjacent to each other are calculated numerically. For example, a pair of partial images that are adjacent to the boundary on the original page of the original document generally have similar patterns on both sides of the boundary, so the power spectrum related to the spatial frequency is used for the likelihood calculation. Can be used.

Specifically, first, for example, as shown in FIG. 9, an area of N pixels wide existing inside the side is divided into N pixel corner blocks. Subsequently, the power spectrum of each block on each side is obtained. That is, the discrete Fourier transform is applied to each block, and the power spectrum (size of spatial frequency component) in each block is obtained. The obtained power spectrum is expressed as Php, i (x, y). Note that hp represents an edge identifier, i represents a block identifier, and (x, y) represents coordinates (spatial frequency) within the block. In addition, when it is determined that all the blocks are background areas such that the pixel values in all corresponding blocks are substantially constant on a certain side, that side (block) is excluded from the calculation target. Then, the likelihood of the combination of the two sides is obtained, for example, according to the following equation (1). In Equation (1), the likelihood value increases as the likelihood that a pair of partial images are adjacent to each other with their sides as boundaries on the original page of the document increases.
Likelihood (hp, hq) = 1 / ΣΣΣ | Php, i (x, y) −Phq, i (x, y) | (1)
ixy

  An example of the likelihood calculation result by the likelihood calculation unit 20 is shown in the likelihood calculation chart of FIG. The combination of sides indicated as “back” in FIG. 10 is a combination of sides determined not to be subjected to likelihood calculation by determining that at least one side is the background region. The likelihood calculation result (likelihood calculation chart) by the likelihood calculation unit 20 is stored in the likelihood storage unit 22. The likelihood calculation method described above is merely an example, and other known calculation methods may be applied.

  In the next step 108, the combination determination unit 24 selects the combination with the maximum likelihood from the combinations of the sides of the partial images that have not been deleted from the likelihood calculation chart. For example, in the likelihood calculation chart shown in FIG. 10, the combination of the side h9 and the side h10 is likelihood = 50, and the likelihood is the maximum combination. In step 114, the combination determination unit 24 saves the combination of the sides selected in step 108 in the combination storage unit 26 as a set of sides to be connected (highly likely to be adjacent on the original page of the original). To do. In Step 116, combination determination part 24 deletes a pair of sides which constitute a combination of sides selected at Step 108 from a likelihood calculation chart, respectively.

  In the next step 118, the combination determination unit 24 determines whether or not a set of all sides of the partial image excluding the side determined as the background region has been determined. If the determination in step 118 is negative, the process returns to step 108, and steps 108 to 118 are repeated until the determination in step 118 is positive. As a result, when the determination in step 118 is affirmed and the image combination processing is terminated, the information shown in FIG. 11 is stored in the combination storage unit 26 as information on the combination of sides to be connected. .

  As described above, in the first embodiment, front / back identification information representing a pair of partial images having a front / back relationship is acquired, and a partial image that is not established as a combination on the original page of the document based on the acquired front / back identification information. Are excluded from the likelihood calculation target. This makes it possible to determine the correct combination of partial images even when there is no clear difference between the image edges of the individual partial images, for example, when the original document is a character original in which characters are written on the entire surface. Accuracy can be improved.

[Second Embodiment]
Next, a second embodiment of the disclosed technique will be described. Since the second embodiment has the same configuration as that of the first embodiment, the same reference numerals are given to the respective parts, and the description of the configuration is omitted. Hereinafter, the first embodiment will be described as the operation of the second embodiment. Only the parts different from the form will be described.

  FIG. 12 shows image combination processing according to the second embodiment. Also in step 102 of the image combination process shown in FIG. 12, the verification information acquisition unit 14 acquires verification information that can determine a combination of partial images that is not established as a combination on the original page of the document. However, in the first embodiment, the front / back identification information is acquired from the scanner 44 as the verification information, but in the second embodiment, the original size information indicating the size of the original page of the original is used as the verification information. Then, partial image size information representing the size of the partial image is acquired.

  For obtaining the document size information, for example, a message requesting the user to input the size of the original page of the document is displayed on the display 38, and the user inputs the document size information via the input device 40. Can be obtained at. On the other hand, the size of the partial image can be calculated from the resolution of document reading by the scanner 44 and the number of pixels of the partial image, and the size of the partial image is recognized by the scanner 44 that has read the partial document. Accordingly, the partial image size information can be acquired from the scanner 44. The obtained document size information and partial image size information are stored in the verification information storage unit 16.

  Further, in the image combination processing according to the second embodiment, the step 104 of determining a combination of sides of partial images that are not established as a combination on the original page of the document and excluding it from the likelihood calculation target is omitted. . Therefore, in the next step 106, the likelihood calculation unit 20 has a positional relationship in which all the combinations of the sides of the partial images are adjacent to each other on the original page on the original page. A likelihood representing the likelihood is calculated. In FIG. 13, an example of the calculation result of the likelihood by the likelihood calculation part 20 in this 2nd Embodiment is shown as a likelihood calculation chart.

  In the next step 108, the combination determination unit 24 selects the combination with the maximum likelihood from the combinations of the sides of the partial images that have not been deleted from the likelihood calculation chart. For example, in the likelihood calculation chart shown in FIG. 13, the combination of the side h9 and the side h10 is likelihood = 50, and the likelihood is the maximum combination. In the next step 110, the exclusion combination determination unit 18 verifies whether or not the combination of the sides of the partial image selected by the combination determination unit 24 is not a combination on the original page of the document.

  Details of the verification process in step 108 will be described with reference to FIG. In step 130 of the combination verification process shown in FIG. 14, the excluded combination determination unit 18 reads the partial image size information from the verification information storage unit 16, and includes the part having the combination verification target side selected by the combination determination unit 24. The sum of the width dimensions of the image is set to the variable w. For example, when a combination of the side h9 and the side h10 is selected by the combination determining unit 24, the sum of the width size of the partial image Image4 having the side h9 and the width size of the partial image Image5 having the side h10 is set to the variable w.

  In the next step 132, the exclusion combination determination unit 18 searches whether there is a pair of sides whose combination has already been determined in the other sides of the partial image having the combination verification target side. In step 134, the exclusion combination determination unit 18 determines whether or not a corresponding side pair has been found in the search in step 132. For example, since the combination of the side h9 and the side h10 is the combination of the sides first selected by the combination determination unit 24, the determination in step 134 is negative and the process proceeds to step 140.

  In step 140, the exclusion combination determination unit 18 reads the document size information from the verification information storage unit 16, and the variable w is equal to or less than the maximum width of the original document obtained from the original page size represented by the read document size information. It is determined whether or not. For example, when the original page size of the document represented by the document size information is A3 size, it is determined whether or not the width dimension set in the variable w is equal to or less than the A3 size width dimension. For example, when the combination of the side h9 and the side h10 is selected by the combination determining unit 24, the sum of the width size of the partial image Image4 having the side h9 and the width size of the partial image Image5 having the side h10 is equal to or less than the width size of the A3 size. Therefore, the determination at step 140 is affirmed and the routine proceeds to step 142.

  In step 142, the exclusion combination determination unit 18 determines that the combination of sides to be verified for combination is established as a combination of the original pages of the original (passes verification), and ends the combination verification processing that does not hold. The process proceeds to step 112 in the flowchart of FIG.

  In step 122, the combination determination unit 24 determines whether the combination of the edges of the previously selected partial image has been determined to be a combination that is not established in the combination verification process that is not established by the exclusion combination determination unit 18 (whether the verification has failed). To determine. If the determination in step 112 is negative, the process proceeds to step 114. In step 114, the combination determination unit 24 saves the combination of the sides selected in step 108 in the combination storage unit 26 as a set of sides to be connected (highly likely to be adjacent on the original page of the original). To do. In Step 116, combination determination part 24 deletes a pair of sides which constitute a combination of sides selected at Step 108 from a likelihood calculation chart, respectively.

  On the other hand, in the likelihood calculation chart shown in FIG. 13, the combination with the maximum likelihood selected next in step 108 is a combination of the side h7 and the side h11 with likelihood = 45. When the combination of the side h7 and the side h11 is selected, the sum of the width size of the partial image Image3 having the side h7 and the width size of the partial image Image5 having the side h11 in the combination verification processing step 130 shown in FIG. Is set in the variable w. In the next step 132, the exclusion combination determination unit 18 searches whether there is a pair of sides whose combination has already been determined in the other sides of the partial image having the sides to be verified for combination. The partial image Image5 is combined with the partial image Image4. Therefore, as a result of the search in step 132, a set of sides h9 and h10 is found as a set of already determined sides, the determination in step 134 is affirmed, and the routine proceeds to step 136.

  In step 136, the exclusion combination determination unit 18 sets the width of another partial image (partial image Image4 in the above example) having one side of the set of determined sides that has been found to the variable v. In step 137, the exclusion combination determination unit 18 adds the determined determined set of sides to the combination verification target side. In the next step 138, the exclusion combination determination unit 18 adds the variable v to the variable w, and returns to step 132. Here, since a set of edges that have already been determined in the search in step 132 is not newly found, the determination in step 134 is negative and the process proceeds to step 140. At this time, the sum of the width dimensions of the partial images Image3, Image5, and Image4 is set in the variable w. Accordingly, in the determination in step 140, the variable w exceeds the maximum width (A3 size width) of the original document, the determination in step 140 is denied, and the process proceeds to step 144.

  In step 144, the exclusion combination determination unit 18 determines that the combination of sides to be verified for combination is not established as a combination on the original page of the original (verification fails). The process proceeds to step 112 in the flowchart of FIG. In this case, the determination in step 112 is affirmed, the process proceeds to step 116, and the pair of sides to be verified for combination is not stored as the set of sides to be connected, but the pair of sides selected in step 108 is formed. Each edge is deleted from the likelihood calculation chart.

  Therefore, although the combination of the side h7 and the side h11 has a relatively high likelihood, it is determined that the combination is not established based on the sum of the width dimensions of the partial images exceeding the maximum width of the original document. It is not saved as a pair of target edges. In the second embodiment, by repeating the above processing, a set of sides to be connected is determined while eliminating combinations of sides where the sum of the width dimensions of the partial images exceeds the maximum width of the original document. .

  As described above, in the second embodiment, original size information and partial image size information are acquired, and combinations of sides of partial images exceeding the maximum width of the original original obtained from the acquired original size information are not combined. Is excluded from the pair of sides to be connected. This makes it possible to determine the correct combination of partial images even when there is no clear difference between the image edges of the individual partial images, for example, when the original document is a character original in which characters are written on the entire surface. Accuracy can be improved.

[Third Embodiment]
Next, a third embodiment of the disclosed technique will be described. Since the third embodiment has the same configuration as the first and second embodiments, the same reference numerals are given to the respective parts, and the description of the configuration is omitted. Hereinafter, the operation of the third embodiment will be described. Only the parts different from the first embodiment will be described.

  In the first embodiment, in step 102 of the image combination process (FIG. 5), the verification information acquisition unit 14 identifies front and back as verification information that can determine a combination of partial images that cannot be established as a combination on the original page of the document. Information was acquired from the scanner 44. On the other hand, in the third embodiment, in step 102 of the image combination process (FIG. 5), the verification information acquisition unit 14 stores the verification information that can determine the combination of the partial images that are not established in the scan image storage unit 12. The top-and-bottom direction of the character included in each individual partial image is detected. The detection of the vertical direction of the character can be realized by applying character recognition processing (OCR: Optical Character Recognition) to the partial image. And the verification information acquisition part 14 acquires the top-and-bottom direction showing the detected top-and-bottom direction for every partial image. Thereby, the top-and-bottom direction information as shown in FIG. 15 is acquired and stored in the verification information storage unit 16.

  Note that the vertical direction of each partial image may be detected by the scanner 44. The top-and-bottom direction information shown in FIG. 15 includes two types of directions that are downward or upward, but may include leftward or rightward directions. In addition, instead of detecting the top and bottom direction of each partial image and recording the top and bottom direction information, the individual partial images are rotated according to the detection result of the top and bottom direction so that the top and bottom directions of all the partial images are aligned. It is also possible to perform the process of saving again. In this case, it is not necessary to store the top and bottom direction information for each partial image.

  In the first embodiment, in step 104 of the image combination process (FIG. 5), the exclusion combination determination unit 18 uses the front / back identification information stored in the verification information storage unit 16 as a combination on the original page of the document. The combination of the sides of the partial image that was not established was judged. On the other hand, in the third embodiment, in step 104 of the image combination process (FIG. 5), the exclusion combination determination unit 18 performs a combination verification process that does not hold as shown in FIG. Is determined as a combination of sides of partial images that do not hold. Hereinafter, the combination verification process which is not realized shown in FIG. 16 will be described.

  In step 150, the exclusion combination determination unit 18 selects a combination of arbitrary sides as a combination of sides to be verified for combination, the top and bottom directions of the pair of partial images having the sides to be verified for combination, and the pair of partial images. Reads the position (left side or right side) of the side for combination verification. In the next step 152, the exclusion combination determination unit 18 determines that the top and bottom direction of the pair of partial images and the position of the combination verification target side in the pair of partial images are both the same, or the top and bottom direction and the side position are both different. It is determined whether or not.

  The condition for combination of partial images that cannot be established as a combination on the original page of the original is that the two sides of the combination verification target are both at the same position (the left sides are the same), even though the vertical direction of the pair of partial images is the same. Or right sides). In addition, as a condition for the combination of partial images that are not established as a combination on the original page of the original, the two sides of the combination verification target are in opposite positions (even though the vertical direction of the pair of partial images is reverse) Left side and right side). That is, if the determination in step 152 is affirmative, it can be determined that a combination that connects a pair of partial images with the two sides to be combined as a boundary is not established.

  Therefore, if the determination in step 152 is affirmed, the process proceeds to step 154. In step 154, the exclusion combination determination unit 18 does not establish the combination of the sides to be verified as a combination on the original page of the document. Judged (failed verification). On the other hand, if the determination in step 152 is affirmed, the process proceeds to step 156. In step 156, the exclusion combination determining unit 18 establishes the combination of the sides to be verified as combinations on the original page of the document ( Judgment is passed).

  In the third embodiment, the exclusion combination determination unit 18 performs the combination verification processing that does not hold as illustrated in FIG. 16 for all combinations of the sides of the partial images. Then, for the combination of sides determined not to be established, the likelihood calculation unit 20 sets the likelihood indicating that the combination is not included in the likelihood calculation chart, as in the first embodiment. Perform processing to be excluded from the target of degree calculation. As described above, prior to the likelihood calculation performed by the likelihood calculation unit 20, the calculation load of the likelihood calculation unit 20 is reduced by excluding the combination in which the side of the partial image is not established from the calculation target. The accuracy of calculation is improved.

  As described above, in the third embodiment, the top / bottom direction information indicating the top / bottom direction of each partial image is acquired, and it is determined that the combination of the top / bottom direction of the pair of partial images and the position of the side in the pair of partial images does not hold. The possible combinations of edges are excluded from the likelihood calculation target. This makes it possible to determine the correct combination of partial images even when there is no clear difference between the image edges of the individual partial images, for example, when the original document is a character original in which characters are written on the entire surface. Accuracy can be improved.

[Fourth Embodiment]
Next, a fourth embodiment of the disclosed technique will be described. Since the fourth embodiment has the same configuration as that of the first to third embodiments, the same reference numerals are given to the respective parts, and the description of the configuration is omitted. Hereinafter, the operation of the fourth embodiment will be described. Only the parts different from the second embodiment will be described.

  In the second embodiment, in step 102 of the image combination process shown in FIG. 12, the verification information acquisition unit 14 acquires document size information and partial image size information as verification information that can determine a combination of partial images that are not established. It was. In the fourth embodiment, when acquiring the verification information in step 102, the verification information acquisition unit 14 performs the verification information acquisition process shown in FIG. Hereinafter, the verification information acquisition process illustrated in FIG. 17 will be described.

  In step 160 of the verification information acquisition process shown in FIG. 17, the verification information acquisition unit 14 displays a page number area designation screen 76 as shown in FIG. 18 on the display 38 as an example, and the page number on the original page of the document is displayed. Requests the user to specify the area (range) described. In the next step 162, the verification information acquisition unit 14 determines whether or not the page number description area is designated by the user, and repeats step 162 until the determination is affirmed. When the page number area designation screen 76 as shown in FIG. 18 is displayed on the display 38, the user operates the input device 40 to display a figure (for example, a rectangular figure) surrounding the area where the page number is described. An operation of designating a page number description area is performed by drawing. In addition, the user specifies that the numerical value existing in the lowermost character string in the partial image is regarded as the page number instead of drawing the figure surrounding the area where the page number is described. There is also an area specification method.

  When the page number description area is designated by the user, the determination at step 162 is affirmed and the routine proceeds to step 164. Note that the range information indicating the range of the page number description area designated by the user is stored in the verification information storage unit 16. In step 164, the verification information acquisition unit 14 performs character recognition processing (OCR) on all partial images stored in the scan image storage unit 12. In step 166, the verification information acquisition unit 14 takes out a single partial image as a processing target, and based on the result of character recognition processing on the processing target partial image, the page specified by the user on the processing target partial image. It is determined whether or not a number exists in the number description area.

  If the determination in step 166 is affirmed, the process proceeds to step 168. In step 168, the verification information acquisition unit 14 displays information indicating that the page number is described as the page number information of the partial image to be processed. The data is stored in the verification information storage unit 16. If the determination in step 166 is negative, the process proceeds to step 170. In step 170, the verification information acquisition unit 14 indicates that the page number is not described as the page number information of the partial image to be processed. The information is stored in the verification information storage unit 16.

  In the next step 172, the verification information acquisition unit 14 determines whether or not page number information has been stored for all partial images stored in the scan image storage unit 12. If the determination in step 172 is negative, the process returns to step 166, and steps 166 to 172 are repeated until the determination in step 172 is affirmed. As a result, page number information is stored in the verification information storage unit 16 for each partial image, as shown in FIG. 19 as an example.

  In the second embodiment, in step 110 of the image combination process shown in FIG. 12, the exclusion combination determination unit 18 performs the combination verification process that does not hold as shown in FIG. In the fourth embodiment, in step 110 of the image combination process, the exclusion combination determination unit 18 performs a combination verification process that does not hold as shown in FIG. Hereinafter, the combination verification process which is not realized shown in FIG. 20 will be described.

  In step 180 of the combination verification process shown in FIG. 20, the exclusion combination determination unit 18 includes information representing page number presence as page number information in the verification information storage unit 16 among the partial images having the sides to be combined verification target. The total number of stored partial images is set in the variable w. In the next step 182, the exclusion combination determination unit 18 determines whether the value of the variable w is 2 or more. If the determination at step 182 is affirmative, the routine proceeds to step 192. In this case, since there are a plurality of partial images in which page numbers are described in the partial images having the sides to be verified for combination, in step 192, the exclusion combination determining unit 18 Judge that the pair of edges is not established (failed verification).

  On the other hand, if the determination in step 182 is negative, the process proceeds to step 184. In step 184, the excluded combination determination unit 18 searches whether there is a set of sides whose combination has already been determined in the other sides of the partial image having the sides to be verified. In the next step 186, the exclusion combination determination unit 18 determines whether or not a corresponding edge pair has been found in the search in step 184. If the determination in step 186 is negative, the process proceeds to step 194. In this case, since there is only one partial image in which the page number is described in the partial images having the combination verification target side, in step 194, the exclusion combination determination unit 18 determines the combination verification target side. It is judged that the pair of is established (passed verification).

  When the determination in step 186 is affirmed, the process proceeds to step 187, and in step 187, the exclusion combination determination unit 18 adds the determined set of determined edges to the combination verification target edge. In the next step 188, the exclusion combination determination unit 18 determines whether or not the page number is described in another partial image having one side of the determined set of sides. If the determination in step 188 is negative, the process returns to step 182. On the other hand, if the determination in step 188 is affirmed, the process proceeds to step 190, and in step 190, the exclusion combination determination unit 18 increments the variable w by 1 and returns to step 182.

  Therefore, when there is a set of sides whose combination has already been determined in the other side of the partial image having the combination verification target side, the combination of the corresponding side is also included in the combination verification target side and has the combination verification target side. It is determined whether or not there is only one page number description location in the entire partial image. If there is only one page number description location, it is determined that the combination of sides to be verified for combination is satisfied (verification has been passed). Is not established (failed verification).

  As described above, in the fourth embodiment, range information representing the range of the page number description area is acquired, and there are two or more page number description locations in the page where numbers are described in the page number description region. The combination of partial images is excluded from the connection targets as a combination of partial images that are not established. This makes it possible to determine the correct combination of partial images even when there is no clear difference between the image edges of the individual partial images, for example, when the original document is a character original in which characters are written on the entire surface. Accuracy can be improved.

  In the fourth embodiment, the page number is applied as an example of a specific character string in the disclosed technology. However, the present invention is not limited to this, and a date, a document creation source, and the like are described instead of the page number. A character string (for example, ABCDE Corporation) may be applied.

  In the fourth embodiment, the aspect in which the verification information acquisition unit 14 performs the character recognition process (OCR) for the partial image is described. However, the present invention is not limited to this, and the character recognition process for the partial image is performed by the scanner 44. Or may be performed by the exclusion combination determination unit 18.

[Fifth Embodiment]
Next, a fifth embodiment of the disclosed technique will be described. Since the fifth embodiment has the same configuration as the first to fourth embodiments, the same reference numerals are given to the respective parts, and the description of the configuration is omitted. Hereinafter, the operation of the fifth embodiment will be described. Only the parts different from the fourth embodiment will be described.

  In the fourth embodiment, in step 102 of the image combination process illustrated in FIG. 12, the verification information acquisition unit 14 performs the verification information acquisition process illustrated in FIG. 17 as verification information that can determine a combination of partial images that are not established. It was. In the fifth embodiment, the original document is a so-called two-up document in which two pages of images are arranged on one page. When the verification information is acquired in step 102, the verification information The acquisition unit 14 performs verification information acquisition processing shown in FIG. Hereinafter, the verification information acquisition process illustrated in FIG. 21 will be described.

  In step 200 of the verification information acquisition process shown in FIG. 21, the verification information acquisition unit 14 displays a page number area designation screen 80 as shown in FIG. 22 on the display 38 as an example, and the page number of the original page of the document is Request the user to specify two or more areas (ranges) that have been described. In the next step 202, the verification information acquisition unit 14 determines whether or not a page number description area is designated by the user, and repeats step 202 until the determination is affirmed. When the page number area designation screen 80 as shown in FIG. 22 is displayed on the display 38, the user operates the input device 40 to enclose a figure (for example, a rectangular shape) surrounding at least two areas where page numbers are described. The operation of designating a plurality of page number description areas is performed.

  When two page number description areas are designated by the user, the determination at step 202 is affirmed and the routine proceeds to step 204. In step 204, the verification information acquisition unit 14 causes the verification information storage unit 16 to store range information indicating the range of the page number description areas at a plurality of locations designated by the user. Thereby, the verification information storage unit 16 stores, for example, range information of the data structure shown in FIG.

  Note that each numerical value of the range information shown in FIG. 23 uses the upper left of the connected partial image as the origin, the positions of the upper left and lower right vertices of the page number description area, and the length of the side of the scanned image after the connection. The ratio is shown as 1. For example, when the connected scan image has a size of horizontal 1500 pixels and vertical 1000 pixels, the serial number “page position 1” is the upper left vertex existing at the position of coordinates (0,900) and the position of coordinates (150,1000). Means that the page number exists in a rectangular area defined by the lower right vertex.

  In the fourth embodiment, in step 110 of the image combination process shown in FIG. 12, the excluded combination determination unit 18 performs the combination verification process that does not hold as shown in FIG. In the fifth embodiment, in step 110 of the image combination process, the excluded combination determination unit 18 performs a combination verification process that does not hold as shown in FIG. Hereinafter, the combination verification process which is not realized as shown in FIG. 24 will be described.

  In step 210 of the combination verification process shown in FIG. 24, the exclusion combination determination unit 18 combines partial images having sides to be combined and generates a scan image in which partial images having sides to be combined are connected. . In the next step 212, the exclusion combination determination unit 18 reads range information representing a plurality of page number description areas stored in the verification information storage unit 16. In the next step 214, the exclusion combination determination unit 18 performs character recognition processing (OCR) on a plurality of page number description areas represented by the range information read in step 212 in the scanned image generated in step 210. To do.

  In the next step 216, the exclusion combination determination unit 18 determines whether or not numbers exist in two places in the page number description area of the scanned image by the character recognition processing in step 214. If the determination in step 216 is negative, the process proceeds to step 220. In this case, since there is no number in at least one of the two page number description areas of the scanned image, in step 220, the exclusion combination determination unit 18 does not establish the combination of sides to be verified (verification ).

  On the other hand, if the determination in step 216 is negative, the process proceeds to step 218. In step 218, the exclusion combination determination unit 18 determines whether or not the numbers respectively present at two locations in the page number description area of the scanned image are consecutive numbers. If the determination in step 218 is negative, the process proceeds to step 220. In this case, since the numbers respectively present in the two positions of the page number description area of the scanned image are not continuous, the exclusion combination determination unit 18 does not establish the combination of sides to be verified in step 220 (verification) ).

  If the determination at step 218 is affirmative, the routine proceeds to step 222. In this case, numbers exist in two places in the page number description area of the scanned image, and numbers exist in two places in the page number description area. For this reason, in step 222, the exclusion combination determination unit 18 determines that the combination of sides to be verified for combination is established (verification has passed).

  As described above, in the fifth embodiment, a number does not exist in at least one of the two page number description areas of an image obtained by combining partial images having a combination verification target side, or the numbers are not continuous. In this case, it is a combination of partial images that do not hold. This makes it possible to determine the correct combination of partial images even when there is no clear difference between the image edges of the individual partial images, for example, when the original document is a character original in which characters are written on the entire surface. Accuracy can be improved.

  In the first and third embodiments, combinations of partial image edges that do not hold are excluded from likelihood calculation targets based on front / back identification information or top / bottom direction information. In the second, fourth, and fifth embodiments, the combination of the sides of the partial images that are not established is excluded from the connection targets based on the document size information and the partial image size information or the range information. However, the disclosed technique is not limited to this, and the contents described in the embodiments may be combined. That is, for example, based on at least one of the front / back identification information and the top / bottom direction information, the combination of the sides of the partial image that does not hold is excluded from the likelihood calculation target, and after calculating the likelihood, the document size information and the partial image size information, Alternatively, a combination of sides of partial images that are not established may be excluded from the connection targets based on the range information.

  In the above description, the computer 30 provided separately from the scanner 44 is described as functioning as an image processing apparatus according to the disclosed technique. However, the present invention is not limited to this, and the scanner 44 itself is an image according to the disclosed technique. You may make it function as a processing apparatus.

  In the above description, the image combination program 46, which is an example of the image processing program in the disclosed technique, is stored (installed) in the storage unit 36 in advance. However, the image processing program in the disclosed technology can be provided in a form recorded on a recording medium such as a CD-ROM or a DVD-ROM.

  All documents, patent applications and technical standards mentioned in this specification are to the same extent as if each individual document, patent application and technical standard were specifically and individually stated to be incorporated by reference. Incorporated by reference in the book.

  Regarding the above embodiment, the following additional notes are disclosed.

(Appendix 1)
An acquisition unit that acquires verification information for determining that a combination is not established as one page of a document when the combination is combined with another partial image from a partial image that is one page of the document when any one is combined (FIG. 1: 14), a determination unit (FIG. 1:18) for determining a combination that is not established as a combination on the original page of the document based on the verification information acquired by the acquisition unit, and an arbitrary one of the plurality of partial images An image processing apparatus including a combination determination unit (FIG. 1: 24) that excludes combinations determined to be not established by the determination unit.

(Appendix 2)
For an arbitrary combination of a plurality of partial images, a likelihood calculating unit (FIG. 1: 20) for calculating the likelihood of being adjacent on the original page of the manuscript is further provided, and the combination determining unit includes the plurality of partial images Based on the likelihood calculated by the likelihood calculating unit, the original page of the manuscript is selected from combinations obtained by excluding the unsatisfied combinations determined by the determining unit from any combination of partial images. The image processing apparatus according to attachment 1, wherein a combination of partial images is determined.

(Appendix 3)
The image processing apparatus according to supplementary note 2, wherein the likelihood calculation unit calculates the likelihood for a combination obtained by excluding the combination that is not established determined by the determination unit from an arbitrary combination of the plurality of partial images.

(Appendix 4)
The individual partial originals obtained by dividing the page of the original into a plurality of pages are read on both front and back sides (FIG. 3A), and front and back identification information for identifying a pair of partial images obtained by reading both front and back sides. (FIG. 3B) is recorded, the acquisition unit acquires the front / back identification information as the verification information, and the determination unit is identified by the front / back identification information acquired as the verification information by the acquisition unit. The image processing apparatus according to any one of Supplementary Note 1 to Supplementary Note 3, wherein the combination of the pair of partial images to be determined is the combination that is not established (FIG. 8).

(Appendix 5)
The acquisition unit acquires document size information indicating the size of the original page of the document and partial image size information indicating the size of a partial image as the verification information, and the determination unit connects a plurality of partial images. A combination of partial images whose partial image size after connection exceeds the original page size represented by the document size information is determined to be a combination that does not hold (FIG. 14). The image processing apparatus according to claim 1.

(Appendix 6)
The vertical direction of characters included in a plurality of partial images obtained by dividing and reading a page of a document is detected, and the vertical direction information (FIG. 15) representing the detected vertical direction is detected for each partial image. The acquisition unit acquires the vertical direction information for each partial image as the verification information, and the determination unit determines the vertical direction of the pair of partial images and the side to be combined in the pair of partial images. A combination of partial images having the same position, or a combination of partial images in which the top and bottom direction of the pair of partial images and the positions of the sides to be combined in the pair of partial images are different from each other is determined as the combination that does not hold (FIG. 16). The image processing apparatus according to any one of appendices 1 to 5.

(Appendix 7)
The acquisition unit acquires, as the verification information, range information indicating a range in which a specific character string is recorded on the original page of the document (FIGS. 17 and 18), and the determination unit acquires a plurality of partial images. When connected, a combination of partial images in which a specific character string does not exist within the range represented by the range information is determined as the combination that does not hold (FIG. 20). Processing equipment.

(Appendix 8)
The image processing apparatus according to appendix 7, wherein the specific character string is a character indicating a page number, a character string indicating a date, or a designated character string.

(Appendix 9)
The specific character string is a character indicating a page number, and the acquisition unit acquires, as the verification information, range information (FIG. 24) indicating a range in which the character indicating the page number is recorded on the original page of the document. (FIGS. 21 and 22), when the determination unit connects a plurality of partial images, the determination unit can detect a partial image in which a character indicating a page number does not exist at an arbitrary position within the range represented by the range information. The image processing apparatus according to appendix 7, wherein a combination and a combination of partial images that are not consecutive page numbers existing in the range represented by the range information are determined as the combinations that are not established (FIG. 24).

(Appendix 10)
When combining any one of the partial images that become one page of the original, when combining with another partial image, the verification information for determining that the combination is not established as one page of the original is acquired. An image processing method comprising: determining a combination that is not established as a combination on the original page of the original, and excluding a combination that is determined not to be established from an arbitrary combination of the plurality of partial images.

(Appendix 11)
For an arbitrary combination of a plurality of partial images, a likelihood obtained by calculating the likelihood of being adjacent on the original page of the document and excluding the determined combination that has not been established from an arbitrary combination of the plurality of partial images The image processing method according to appendix 10, wherein a combination of partial images to be an original page of the original is determined based on the calculated likelihood.

(Appendix 12)
The image processing method according to claim 11, wherein the likelihood is calculated for a combination obtained by excluding the determined combination that is not established from an arbitrary combination of the plurality of partial images.

(Appendix 13)
Each partial original obtained by dividing the page of the original into a plurality of pages is read on both front and back sides, and front and back identification information for identifying a pair of partial images obtained by reading the front and back sides is recorded. 13. The supplementary note 10 to the supplementary note 12, wherein the combination of the pair of partial images identified by the front and back identification information acquired as the verification information is acquired as the combination that is not established. Image processing method.

(Appendix 14)
As the verification information, original size information indicating the original page size of the original and partial image size information indicating the size of the partial image are acquired, and the partial image size after connecting a plurality of partial images is 14. The image processing method according to any one of supplementary notes 10 to 13, wherein a combination of partial images exceeding a size of an original page of the document represented by the document size information is determined as the combination that is not established.

(Appendix 15)
The vertical direction of characters included in a plurality of partial images obtained by dividing the original page into a plurality of parts is detected, and vertical direction information indicating the detected vertical direction is recorded for each partial image. The vertical direction information for each partial image is acquired as the verification information, and the vertical direction of the pair of partial images and the position of the combination target side in the pair of partial images are both the same, or the pair of partial images The image processing method according to any one of supplementary notes 10 to 14, wherein a combination of partial images in which the positions of the sides to be combined in the top and bottom direction and the pair of partial images are different from each other is determined as the combination that is not established.

(Appendix 16)
As the verification information, range information representing a range in which a specific character string is recorded on the original page of the document is acquired, and when a plurality of partial images are connected, a specific character within the range represented by the range information is obtained. The image processing method according to any one of Supplementary Note 10 to Supplementary Note 15, wherein a combination of partial images in which a combination of partial images in which no column exists is determined as the combination that is not established.

(Appendix 17)
The image processing method according to claim 16, wherein the specific character string is a character indicating a page number, a character string indicating a date, or a designated character string.

(Appendix 18)
The specific character string is a character indicating a page number, and as the verification information, range information indicating a range in which the character indicating the page number is recorded on the original page of the document is acquired, and a plurality of partial images are connected. Sometimes, a combination of partial images in which characters indicating page numbers do not exist at arbitrary locations within the range indicated by the range information, and partial images in which page numbers existing in the range indicated by the range information are not continuous. The image processing method according to supplementary note 16, wherein a combination is determined to be a combination that is not established.

(Appendix 19)
Acquired verification information from a partial image that becomes one page of the original when it is combined with another computer, and verifies that the combination does not hold as one page of the original when combined with another partial image. Based on the verification information, a process including determining a combination that is not established as a combination on the original page of the document and excluding a combination that is determined not to be established from an arbitrary combination of the plurality of partial images is performed. An image processing program for recording or a recording medium on which the image processing program is recorded.

(Appendix 20)
For an arbitrary combination of a plurality of partial images, a likelihood obtained by calculating the likelihood of being adjacent on the original page of the document and excluding the determined combination that has not been established from an arbitrary combination of the plurality of partial images 20. The image processing program or recording medium according to appendix 19, wherein a combination of partial images to be an original page of the original is determined based on the calculated likelihood.

(Appendix 21)
The image processing program or recording medium according to appendix 20, wherein the likelihood is calculated for a combination obtained by excluding the determined combination that is not established from an arbitrary combination of the plurality of partial images.

(Appendix 22)
Each partial original obtained by dividing the page of the original into a plurality of pages is read on both front and back sides, and front and back identification information for identifying a pair of partial images obtained by reading the front and back sides is recorded. The supplementary note 19 to the supplementary note 21, wherein the combination of the pair of partial images identified by the front and back identification information acquired as the verification information is determined as the combination that is not established. Image processing program or recording medium.

(Appendix 23)
As the verification information, original size information indicating the original page size of the original and partial image size information indicating the size of the partial image are acquired, and the partial image size after connecting a plurality of partial images is 23. The image processing program or the recording medium according to any one of supplementary notes 19 to 22, wherein a combination of partial images exceeding a size of an original page of the document represented by the document size information is determined as the combination that is not established.

(Appendix 24)
The vertical direction of characters included in a plurality of partial images obtained by dividing the original page into a plurality of parts is detected, and vertical direction information indicating the detected vertical direction is recorded for each partial image. The vertical direction information for each partial image is acquired as the verification information, and the vertical direction of the pair of partial images and the position of the combination target side in the pair of partial images are both the same, or the pair of partial images 24. The image processing program or recording according to any one of supplementary notes 19 to 23, wherein a combination of partial images in which the positions of sides to be combined in the top-bottom direction and the pair of partial images are different from each other is determined to be a combination that does not hold Medium.

(Appendix 25)
As the verification information, range information representing a range in which a specific character string is recorded on the original page of the document is acquired, and when a plurality of partial images are connected, a specific character within the range represented by the range information is obtained. 25. The image processing program or the recording medium according to any one of supplementary notes 19 to 24, wherein a combination of partial images in which a combination of partial images in which no column exists is determined to be a combination that does not hold.

(Appendix 26)
The image processing program or recording medium according to attachment 25, wherein the specific character string is a character indicating a page number, a character string indicating a date, or a designated character string.

(Appendix 27)
The specific character string is a character indicating a page number, and as the verification information, range information indicating a range in which the character indicating the page number is recorded on the original page of the document is acquired, and a plurality of partial images are connected. Sometimes, a combination of partial images in which characters indicating page numbers do not exist at arbitrary locations within the range indicated by the range information, and partial images in which page numbers existing in the range indicated by the range information are not continuous. The image processing program or recording medium according to appendix 25, wherein a combination is determined to be a combination that is not established.

DESCRIPTION OF SYMBOLS 10 Image processing apparatus 12 Scanned image storage part 14 Verification information acquisition part 16 Verification information storage part 18 Exclusion combination judgment part 20 Likelihood calculation part 22 Likelihood storage part 24 Combination determination part 26 Combination storage part 30 Computer 32 CPU
34 Memory 36 Storage unit 38 Display 40 Input device 44 Scanner 46 Image combination program

Claims (10)

An acquisition unit for acquiring verification information for determining that the combination is not established as one page of a document when combining with any other partial image from the partial image that becomes one page of the document when any one is combined;
A determination unit that determines a combination that is not established as a combination on the original page of the document based on the verification information acquired by the acquisition unit;
A combination determination unit that excludes combinations determined not to be established by the determination unit among arbitrary combinations of the plurality of partial images;
An image processing apparatus. For an arbitrary combination of a plurality of partial images, further comprising a likelihood calculating unit that calculates the likelihood of being adjacent on the original page of the original,
The combination determination unit is based on the likelihood calculated by the likelihood calculation unit from a combination obtained by excluding the unsatisfied combination determined by the determination unit from any combination of the plurality of partial images. The image processing apparatus according to claim 1, wherein a combination of partial images to be an original page of the original is determined.   The image processing device according to claim 2, wherein the likelihood calculation unit calculates the likelihood for a combination obtained by excluding the combination that is not established determined by the determination unit from an arbitrary combination of the plurality of partial images. Front and back both sides of each of the partial originals obtained by dividing the page of the original are read, and front and back identification information for identifying a pair of partial images obtained by reading the front and back sides is recorded.
The acquisition unit acquires the front and back identification information as the verification information,
4. The determination unit according to claim 1, wherein the determination unit determines that the combination of the pair of partial images identified by the front and back identification information acquired as the verification information by the acquisition unit is the combination that is not established. The image processing apparatus according to claim 1. The acquisition unit acquires document size information indicating a size of an original page of the document and partial image size information indicating a size of a partial image as the verification information,
The determination unit is a combination of partial images in which a partial image size after connecting a plurality of partial images exceeds a size of an original page of the document represented by the document size information is a combination that is not established. The image processing device according to claim 1, wherein the image processing device is determined. The vertical direction of characters included in a plurality of partial images obtained by dividing and reading a page of a document is detected, and vertical direction information indicating the detected vertical direction is recorded for each partial image,
The acquisition unit acquires the vertical direction information for each partial image as the verification information,
The determination unit has the same vertical direction of the pair of partial images and the position of the combination target side in the pair of partial images, or the vertical direction of the pair of partial images and the position of the combination target side in the pair of partial images. The image processing apparatus according to claim 1, wherein a combination of partial images having different values is determined as the combination that is not established. The acquisition unit acquires, as the verification information, range information indicating a range in which a specific character string is recorded on the original page of the document,
The determination unit, when connecting a plurality of partial images, determines a combination of partial images in which a specific character string does not exist within the range represented by the range information as the combination that does not hold. The image processing apparatus according to claim 1. The specific character string is a character indicating a page number,
The acquisition unit acquires, as the verification information, range information indicating a range in which characters indicating page numbers are recorded on the original page of the document,
When the determination unit connects a plurality of partial images, a combination of partial images in which a character indicating a page number does not exist at an arbitrary position within the range represented by the range information, and the range information The image processing apparatus according to claim 7, wherein a combination of partial images that do not have consecutive page numbers within a range to be represented is determined as the combination that is not established. When combining any one, from the partial image that becomes one page of the document, when combining with another partial image, to obtain verification information to determine that the combination is not established as one page of the document,
Based on the acquired verification information, determine a combination that is not established as a combination on the original page of the document,
An image processing method including excluding a combination determined not to be established from an arbitrary combination of the plurality of partial images. On the computer,
When combining any one, from the partial image that becomes one page of the document, when combining with another partial image, to obtain verification information to determine that the combination is not established as one page of the document,
Based on the acquired verification information, determine a combination that is not established as a combination on the original page of the document,
An image processing program for performing processing including excluding a combination determined not to be established from an arbitrary combination of the plurality of partial images.