JP5190312B2 - Seal verification device - Google Patents

Description

  The present invention relates to a seal stamp collation apparatus, and more particularly to a seal stamp collation apparatus capable of accurately discriminating similar characters.

  In the collation process between the registration mark and the reading mark read from the form or the like by the scanner, first, image data obtained by superimposing the registration mark and the reading mark is created, and the matching rate is determined based on the image data.

If the matching rate is lower than a predetermined threshold value, the reading mark does not match the registration mark, and the matching process ends.
On the other hand, when the coincidence rate is higher than a predetermined threshold, a similar character determination process (local analysis process) for determining whether or not the reading mark is a similar character to the registration mark is performed. As a result of this processing, if it is determined that the character is similar, the reading mark eventually does not match the registration mark, and the matching process ends.

As a result of this process, if it is not determined that the character is similar, the read mark finally matches the registration mark, and the collation process ends.
However, conventionally, in the local analysis processing, a uniform parameter (rectangular region size used for determination) is used regardless of the size of the seal impression. This method has a problem that similar character determination may not be performed accurately depending on the size of the seal impression and the size of the difference. Note that similar characters are similar but different characters such as “large” and “thick”.

As a peripheral technique, for example, Patent Document 1 discloses a technique of reading with increased accuracy according to the size of a seal. Patent Document 2 discloses a technique for determining the number of divisions according to the size of the collation area. Moreover, in patent document 3, the image data acquisition part which acquires the image of the stamped imprint surface, the extraction part which extracts the image data which has the color of a collation target impression from the acquired image data, The extracted image data 1 shows a seal collation apparatus having a collation target specifying unit that extracts image data in a range having a size of a collation source seal stamp as image data of a collation target seal stamp.
JP 58-223878 A JP-A-61-182180 Japanese Patent Laid-Open No. 11-96362

  SUMMARY OF THE INVENTION An object of the present invention is to provide a seal collation apparatus that can distinguish similar characters with high accuracy.

  The proposed seal collation apparatus includes a reading unit that reads a seal imprint stamped on a medium as read stamp image data, registered seal image data that is image data of a registered seal stamp, and the read read stamp A registration processing unit that calculates a matching rate between the registered stamp image data and the read stamp image data by superimposing image data, and the registered stamp image when the calculated matching rate is equal to or greater than a threshold value. The fixed-size rectangular area is shifted by one pixel from the upper left end of the image to the lower right end of the image on the image obtained by superimposing the data and the reading mark image data. When the number of different pixels is counted and the number of different pixels is greater than or equal to the threshold value in a rectangular area of any fixed size, A similar character determination unit that determines data as a similar character, and a collation processing result output that outputs a determination result as to whether the registered seal image data and the read seal image data match based on the determination result of the similar character determination unit Part.

  Then, the storage unit that associates the registration mark size or size range with the fixed size of the rectangular area and the threshold value of the number of different pixels is searched using the registered mark size as a key, and the fixed size and difference of the found rectangular area are searched. The determination process by the similar character determination unit is executed using a threshold value of the number of pixels.

  According to the proposed seal stamp collation apparatus, the fixed size of the rectangular area and the threshold value of the number of different pixels are determined according to the size of the seal, and similar character determination processing is performed using the fixed size, which is different from the registered seal. It is possible to reduce the possibility that it is determined that the reading marks match, and to reduce the possibility that the reading mark that is the same as the registered mark is determined to be a similar character. It was discovered by the inventors of the present application through a collation experiment between data. At this time, the ratio of the threshold value of the number of different pixels to the number of pixels included in the filter is substantially constant regardless of the registration mark size, and the value can be selected from the range of 0.877 to 0.89. It has also been discovered by the inventors of the present application that it is preferable.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
Instead of using an automatic transaction device (ATM) when doing various transactions such as deposits, withdrawals, transfers, etc. at a store such as a bank, a notification stamp is stamped on the form corresponding to the transaction and this is the window The case of submitting to is handled below.

  A seal stamp verification device (hereinafter simply referred to as “verification device”) is installed at the counter, and the operator can operate it to verify whether the stamped seal matches the registered one. Is called.

FIG. 1 is a block diagram showing a configuration of a seal collator according to an embodiment of the present invention.
In FIG. 1, the window operator refers to the submitted form and, for example, inputs the account number on the form via the input unit (not shown) of the seal collation apparatus 1. Then, using the account number input by the search unit 2 as a key, the seal server 3 that holds data in which the account number is associated with the image data of the seal stamp registered (hereinafter referred to as “registered seal image data”). Search and read the registered seal image data found from the seal server 3.

Subsequently, the window operator reads a predetermined range on the form including the seal of the seal stamped on the form submitted by the scanner 4 as image data.
Collation / determination is performed in which the image data of the registration mark and the reading mark is collated, and further, it is determined whether or not they are similar characters, and finally the image data of the registration mark and the reading mark is matched. The unit 5 includes an imprint search unit 6, a collation processing unit 7, and a similar character determination unit 8.

  The imprint search unit 6 searches for a predetermined range of image data on the form including the read imprint, and the range of the predetermined range where the imprint exists is read from the image data of the imprint read from the form ( Hereinafter, it is specified as “reading stamp image data”.

The collation processing unit 7 creates registration image data by superimposing the registration mark image data and the reading mark image data, and calculates a matching rate between the registration mark image data and the reading mark image data based on the overlapped image data. Note that, when performing superposition, the size (longer length in the vertical / horizontal direction) of the seal area of the registration mark is obtained from the registration mark image data.

  The similar character determination unit 8 performs a determination process (in order to exclude a reading mark similar to the registration mark when the matching rate between the registration mark image data and the reading mark image data is equal to or higher than a predetermined threshold value ( Details will be described later).

  Note that if the matching rate between the registered mark image data and the read mark image data is smaller than a predetermined threshold, it is clear that the registered mark image data and the read mark image data do not match. Similar character determination processing is omitted. Note that “similar characters” mean similar but different characters.

  The collation result display unit 11 determines whether the registered stamp image data and the read stamp image data match based on the combination of the processing result of the collation processing unit 7 and the determination result of the similar character determination unit 8, and the determination result Is output on the display device 12.

Next, similar character determination processing (local analysis processing) will be described with reference to FIGS. 2A, 2B, and 3. FIG.
2A and 2B are views (No. 1) illustrating a method of similar character determination processing (local analysis processing).

  Here, a case where the registration mark is “Ota” and the reading mark read from the form is “Ota” is handled. In this case, as shown in FIG. 2A, after the registration mark image data and the reading mark image data are overlapped to generate the overlap image data 15, the overlap image data (difference portion pixels) 16 obtained by extracting the different portions of the pixels. Create

That is, each pixel of the superimposed image data 15 has a value that identifies a combination of a corresponding pixel value of the registration mark image data and a corresponding pixel value of the reading mark image data. An example is shown below.
(1) When the pixels of the corresponding read mark image data and registration mark image data are both black, the pixel of the superimposed image data is red.
(2) When the corresponding pixel of the read mark image data is black and the corresponding pixel of the registered mark image data is white, the pixel of the superimposed image data is blue.
(3) When the pixel of the corresponding read mark image data is white and the pixel of the corresponding registered mark image data is black, the pixel of the superimposed image data is green.
(4) When the pixels of the corresponding read mark image data and registration mark image data are both white, the pixels of the superimposed image data are white.

The “difference part of the pixel” means the parts (2) and (3).
The similar character determination process may be performed on the superimposed image data 15 having the four colors (1) to (4), or the superimposed image data (differences) having the two colors (2) and (3). (Partial pixels) 16 may be performed.

In the present embodiment, a filter 18 (also referred to as a “square”) having a size corresponding to the registered mark size or size range is used.
In FIG. 2B (1), this filter 18 is shifted by one dot from the upper left end of the superimposed image data (different partial pixel) 16 to the right end, and the number of different pixels in any of the filters 18 (registered mark size). Alternatively, it is detected whether or not the threshold is greater than or equal to the size range.

  When the cell 18 reaches the right end, as shown in (2) of FIG. 2B, the cell returns to the left end and the 1-dot cell 18 is shifted downward. Then, as shown in (3), at the position shifted downward by one dot, the grid 18 is similarly shifted by one dot to the right end. In this way, the process is repeated, and as shown in (4), a different portion is detected from the entire image.

  In actual processing, if it is determined that the number of different pixels is greater than or equal to the threshold at any position of the filter (mesh) 18, the read mark image data and the registered mark image data are determined to be similar characters. Is done.

  For example, for the filter 18 in the superimposed image data 15 of FIG. 3, the size of the filter 18 is 7 × 7 = 49 pixels, of which 43 are the number of different pixels, and the threshold for the number of different pixels is Since it is 43, the number of different pixels matches the threshold value and is determined to be a similar character.

Next, the current problem of the similar character determination process (local analysis process) will be described with reference to FIGS.
In the current similar character determination process, the similar character determination parameters are uniformly set regardless of the size of the seal. That is, regardless of the registration mark size, the size of the filter 18 is 7 × 7 = 49 pixels, and the threshold value of the number of different pixels is 43.

FIG. 4 is a diagram illustrating an example of a processing result of the current local analysis processing for a standard size imprint (difference stamp).
In the example of FIG. 4, the registration mark image data 21 has a size of vertical = about 1.0 cm and horizontal = about 1.0 cm, and from this, the size of the imprint area of the registration mark (the longer length in the vertical / horizontal direction). ) Is about 1.0 cm. The read mark image data 22 is clearly different from the registered mark image data 21 in the example of FIG. Since the number of different pixels is 43 and the threshold value of the number of different pixels is 43 for the filter 24 in the superimposed image data 23 obtained by superimposing these, it is determined that they are similar characters. That is, there is no problem with a standard size imprint.

FIG. 5 is a diagram illustrating an example of a processing result of the current local analysis processing for a small imprint (difference stamp) (first case in which the determination result is to be corrected).
In the example of FIG. 5, the registration mark image data 26 has a size of vertical = about 0.7 cm and horizontal = about 0.7 cm, and from this, the size of the imprint area of the registration mark (the longer length in the vertical / horizontal direction). ) Is about 0.7 cm. The read mark image data 27 is clearly different from the registered mark image data 26 in the example of FIG. For the filter 29 in the superimposed image data 28 obtained by superimposing these, the number of different pixels is 40 (≦ threshold 43), and it is not determined that the character is similar. That is, the registration mark and the reading mark are clearly different, but the imprint is small, so the difference is relatively small compared to the standard size. For this reason, although it should determine with a similar character, it is not determined.

FIG. 6 is a diagram illustrating an example of a processing result of the current local analysis processing for a large imprint (same mark) (second case in which the determination result is to be corrected).
In the example of FIG. 6, the registration mark image data 31 has a size of vertical = about 1.7 cm and horizontal = about 1.7 cm, and from this, the size of the imprint region of the registration mark (the longer length in the vertical / horizontal direction). ) Is about 1.7 cm. The read mark image data 32 is the same as the registered mark image data 26 in the example of FIG. 6, but when the character density is high as in the case of the corporate mark in this case, the number of different pixels increases due to the blurring of characters. Yes. For the filter 34 in the superimposed image data (different pixel portion) 33 obtained by superimposing these, the number of different pixels is 48 (≧ threshold value 43), and it is determined that they are similar characters. That is, although the registration mark and the reading mark are the same mark, the character density is high and there are many different pixels due to the bleeding of the character.

FIG. 7 is a table in which the registered mark size range and the similar character discrimination parameter are associated with each other.
In the table of FIG. 7, the filter size is 5 × 5 = 25 pixels, different pixels for a registration mark having a small size (registration marks of 0.3 cm to 0.9 cm, registration marks less than 0.3 cm are unlikely). The number threshold is set to 22, and the size of a standard registration mark (registration mark of 0.9 cm to 1.3 cm) is set to 7 × 7 = 49 pixels, the difference pixel number threshold is set to 43, and the size is set. For large registration marks (registration marks of 1.3 cm to 5.0 cm, registration marks larger than 5.0 cm are unlikely), the filter size is 10 × 10 = 100 pixels and the threshold for the number of different pixels is 89 .

  In the above table, the size of the seal is divided into three sections, but any number other than three can be used. Also, values other than those described above can be adopted for the filter size and the threshold value for the number of different pixels set for each section.

  By using the table shown in FIG. 7 to determine the filter size and the threshold value of the number of different pixels according to the size of the seal impression, and using them to perform similar character determination processing, for example, the overlap shown in FIG. As shown in FIG. 8, the filter 36 in the combined image data 28 has 25 (≧ 22) different pixels and is determined to be a similar character. Further, as shown in FIG. 9, the filter 38 in the superimposed image data (different pixel portion) 33 shown in FIG. 6 has 63 (≦ 89) different pixels and is not determined to be a similar character. .

  As described above, according to this embodiment, the filter size and the threshold value of the number of different pixels are determined according to the size of the seal impression, and similar character determination processing is performed using the filter size and the difference between the registered seal and the reading. It is possible to reduce the possibility that the marks are determined to match, and to reduce the possibility that the reading mark that is the same as the registered mark is determined to be a similar character. It was discovered by the inventor of the present application through a collation experiment (parts of which are shown in FIGS. 3, 4, 5, 6, 8, and 9 in the present application). At this time, the ratio of the threshold value of the number of different pixels to the number of pixels included in the filter is substantially constant regardless of the registration mark size, and the value is set to 0.877 (= 43/49) to 0. It was also discovered by the inventors of the present application that it is preferable to select from the range of 89 (89/100).

FIG. 10 is a flowchart of the seal verification process.
In step S11 of FIG. 10, the submitted form is referred to by the operator at the window, and for example, the account number on the form is input via the input unit (not shown) of the seal collation apparatus 1. Then, the search server 2 searches the seal server 3 holding data in which the account number and the registered seal image data are associated with each other using the input account number as a key, and the registered server image data found is read from the seal server 3. It is.

  Subsequently, in step S12, the operator at the window reads the predetermined range on the form including the seal of the seal stamped on the form submitted by the scanner 4 as image data.

  In step S13 subsequent to step S12, the imprint search unit 6 searches for a predetermined range of image data on the form that includes the read imprint, and the range of the predetermined range that includes the imprint is read image. Identify as data.

  In step S14 following step S13, the collation processing unit 7 collates the registered mark image data and the read mark image data, and as a result, the matching rate between the registered mark image data and the read mark image data is calculated. The At this time, the size (longer length in the vertical / horizontal direction) of the seal area of the registration mark is obtained from the registration mark image data.

In subsequent step S15, the registration mark image data and the reading mark image data are superimposed to create a superimposed image data.
In step S16 following step S15, parameters for similar character determination, that is, the table of FIG.

  In step S17, the size of the imprint area of the registration mark obtained in step S14 is acquired. In step S18, the size of the imprint area of the registration mark is determined, and the process branches according to the size.

  That is, when it is determined in step S18 that the size of the registration mark is included in the size range of 0.3 cm to 0.9 cm, in step S19, the corresponding item in the table of FIG. X5 = 25 pixels, the threshold value of the number of different pixels 22) is set as a parameter for similar character determination, and the similar character determination processing in step S22 is executed by the similar character determination unit 8 using the parameter.

  If it is determined in step S18 that the size of the registration mark is included in the size range of 0.9 cm to 1.3 cm, the corresponding item (filter size is set to 7 in the table of FIG. 7) in step S20. X7 = 49 pixels, the threshold value of the number of different pixels is set as 43) as a parameter for determining similar characters, and the similar character determining process in step S22 is executed by the similar character determining unit 8 using the parameters.

  If it is determined in step S18 that the size of the registration mark is included in the size range of 1.3 cm to 5.0 cm, in step S21, the corresponding item (filter size is set to 10 in the table of FIG. 7). × 10 = 100 pixels, the threshold value of the number of different pixels 89) is set as a parameter for similar character determination, and the similar character determination processing in step S22 is executed by the similar character determination unit 8 using the parameter.

  The similar character determination unit 8 sets the similar character determination flag to ON when the number of different pixels in the filter exceeds the threshold at any position. If there is no such position, the process is continued with the similar character discrimination flag off until the last position.

  In step S23 subsequent to step S22, it is determined whether there is a filter position in which the number of different pixels in the filter is equal to or greater than the threshold value of similar characters (threshold value of different pixels) according to the value of the similar character determination flag.

  If it is determined in step S23 that there is a filter position equal to or greater than the threshold value for similar characters (if the similar character determination flag is on, the determination result in step S23 is Yes), the calculation is performed in step S14 in step S14. The penalty rate is corrected by subtracting the penalty point from the match rate. This is, for example, a significant subtraction so that the registration mark image data and the reading mark image data do not always match. As a result, in the subsequent step S25, the result of the collation process that the registered mark image data and the read mark image data do not match is displayed on the display device 12.

  On the other hand, if it is determined in step S23 that there is no filter position equal to or greater than the threshold value for similar characters (if the similar character determination flag is off, the determination result in step S23 is No), in the subsequent step S25, the registration mark image is displayed. The result of the collation process that the data and the reading mark image data match is displayed on the display device 12.

Actually, after the match rate is calculated in step S14, it is determined whether or not the similar character determination process is performed by determining whether or not the match rate is equal to or higher than a threshold (match rate). There is. In this step, when the matching rate exceeds the threshold, the similar character determination process is executed.

  In the above description, in step S24, the penalty point is subtracted to correct the matching rate and reflect it in the final result of the matching process. However, other methods can be adopted. For example, the matching rate may be determined by a combination of the matching rate and the similar character determination flag without correcting the matching rate. In this case, if the matching rate is equal to or higher than the threshold value and the similar character determination flag is on, it is determined that the registered mark image data and the read mark image data do not match, the matching rate is equal to or higher than the threshold value, and the similar character determination flag is off. In this case, it is determined that the registered stamp image data and the read stamp image data match.

Subsequently, similar character determination processing (local analysis processing) will be described with reference to FIGS. 11 and 12.
FIG. 11 is a flowchart of similar character determination processing (local analysis processing).

  In this similar character determination process, as shown in FIG. 12, the set filter 43 is moved from the upper left end of the superimposed image data 42 to the lower right end while shifting one dot at a time. The processing in steps S31 to S38 in FIG. 11 is executed in each set filter. In counting the number of different pixels, in the above (2), when the pixel of the corresponding read mark image data is black and the corresponding pixel of the registered mark image data is white, the pixel of the superimposed image data is blue. In the case where the pixel of the corresponding read mark image data is white and the pixel of the corresponding registered mark image data is black in the above (3), the pixel of the superimposed image data is green Count cases independently.

  Prior to the processing in step S31 of FIG. 11, the vertical pixel range and horizontal pixel range of the rectangular region of the filter are set. Thereby, the position of the filter 43 on the superimposed image data 42 is determined.

In step S31 of FIG. 11, a count variable for the number of blue pixels, a count variable for the number of green pixels, and a count variable for the number of blue or green pixels are initialized.
In subsequent step S32, it is determined whether or not the target pixel (current pixel) is within the filter size region.

  If it is determined in step S32 that the pixel of interest is within the filter size region (if the determination result in step S32 is Yes), if the value of the pixel of interest is blue, a blue pixel is detected in steps S33 and S35. The count variable for the number and the count variable for the number of blue or green pixels are respectively incremented, and the process returns to step S32 with the next pixel as the target pixel.

  If the value of the target pixel is green, the count variable for the number of green pixels and the count variable for the number of blue or green pixels are incremented in steps S34 and S35, respectively, and the next pixel is set as the target pixel. Return to S32.

  If it is determined in step S32 that the target pixel is not in the filter size area (if the determination result in step S32 is No), the number of blue pixels in the filter area on the superimposed image data 42 is determined in step S36. It is determined whether the count variable is less than the threshold value of the number of different pixels.

If it is determined in step S36 that the count variable for the number of blue pixels is not less than the threshold (that is, greater than or equal to the threshold) (if the determination result in step S36 is No), the similar character determination flag is turned on in step S39. To complete the series of processing.

  If it is determined in step S36 that the count variable for the number of blue pixels is less than the threshold value (if the determination result in step S36 is Yes), in step S37, the green color in the filter area on the superimposed image data 42 is determined. It is determined whether the pixel count variable is less than the difference pixel threshold.

  If it is determined in step S37 that the count variable for the number of green pixels is not less than the threshold (that is, greater than or equal to the threshold) (if the determination result in step S37 is No), the similar character determination flag is turned on in step S39. To complete the series of processing.

  If it is determined in step S37 that the count variable for the number of green pixels is less than the threshold value (if the determination result in step S37 is Yes), in step S38, the blue color in the filter area on the superimposed image data 42 or It is determined whether the count variable for the number of green pixels is less than the threshold for the number of different pixels.

  If it is determined in step S38 that the count variable for the number of blue or green pixels is not less than the threshold (that is, greater than or equal to the threshold) (if the determination result in step S38 is No), a similar character determination flag is determined in step S39. Is set to ON, and a series of processing ends.

  When it is determined in step S38 that the count variable for the number of blue or green pixels is less than the threshold (when the determination result in step S38 is Yes), the vertical pixel range and the horizontal pixel range of the rectangular region of the filter Is set to the following, and the processing after step S31 is repeated.

It is a block diagram which shows the structure of the seal | sticker collation apparatus which concerns on one Embodiment of this invention. It is FIG. (1) which shows the method of a similar character determination process (local analysis process). It is a figure (the continuation of the 1) which shows the method of a similar character determination process (local analysis process). It is FIG. (2) which shows the method of a similar character determination process (local analysis process). It is a figure which shows an example of the process result of the present local analysis process in a standard size imprint (difference mark). It is a figure which shows an example of the process result of the present local analysis process in a small imprint (difference mark). It is a figure which shows an example of the process result of the present local analysis process in a big imprint (same mark). It is the table which matched the size of the registration mark and the parameter of similar character discrimination. It is a figure which shows an example of the process result of the local analysis process in the small imprint (difference mark) by the method of this embodiment. It is a figure which shows an example of the process result of the local analysis process in the big imprint (same mark) by the method of this embodiment. It is a flowchart of a seal | sticker collation process. It is a flowchart of a similar character determination process (local analysis process). It is FIG. (The 3) which shows the method of a similar character determination process (local analysis process).

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Seal verification apparatus 2 Search part 3 Seal server 4 Scanner 5 Collation / determination part 6 Imprint search part 7 Collation process part 8 Similar character determination part 11 Collation result display part 12 Display apparatus 15,23,28,42 Superimposition image data 16 , 33 Overlaid image data (different pixel)
18, 24, 29, 34, 36, 38, 43 Filter 21, 26, 31 Registration mark image data 22, 27, 32 Reading mark image data 40 Environment file

Claims (6)

A reading unit that reads a seal stamp of a seal stamped on a medium as read seal image data;
Collation for calculating the coincidence ratio between the registered stamp image data and the read stamp image data by superimposing the registered stamp image data, which is the image data of the registered seal stamp, and the read stamp image data. A processing unit;
When the calculated coincidence rate is equal to or greater than a threshold value, a fixed-size rectangular area is shifted from the upper left end of the image to the lower right end of the image on the image obtained by superimposing the registered mark image data and the read mark image data. Each time the pixel is shifted, the number of different pixels in the fixed-size rectangular area is counted, and if the number of different pixels is greater than or equal to the threshold value in any fixed-sized rectangular area, the reading is performed. A similar character determination unit for determining the stamp image data as a similar character;
A collation processing result output unit that outputs a determination result as to whether the registration mark image data and the read mark image data match based on the determination result of the similar character determination unit;
The storage unit that associates the registered mark size or size range with the fixed size of the rectangular area and the threshold value of the number of different pixels is searched using the registered mark size as a key, and the fixed size and the number of different pixels of the found rectangular area The seal stamp collating apparatus is characterized in that the determination process by the similar character determination unit is executed using the threshold value.   In the storage unit, a ratio of the threshold value of the number of different pixels to the number of pixels included in the fixed-size rectangular area is substantially constant regardless of the registration mark size, and the value is 0.877 to 0.89. The seal stamp verification apparatus according to claim 1, wherein the seal stamp verification apparatus is provided.   One of the fixed sizes of the rectangular area is a rectangular area of 5 × 5 25 pixels, and the threshold of the number of different pixels is 22 pixels with respect to the rectangular area of 5 × 5 25 pixels. The seal collation apparatus according to claim 2.   One of the fixed sizes of the rectangular area is a 7 × 7, 49-pixel rectangular area, and the threshold of the number of different pixels is 43 pixels with respect to the 7 × 7, 49-pixel rectangular area. The seal collation apparatus according to claim 2.   One of the fixed sizes of the rectangular area is a rectangular area of 10 × 10 100 pixels, and the threshold of the number of different pixels is 89 pixels with respect to the rectangular area of 10 × 10 100 pixels. The seal collation apparatus according to claim 2. The collation result output unit determines that the registered stamp image data and the read stamp image data match when the similar character determination unit determines that the character is not similar,
The seal stamp verification apparatus according to claim 1, wherein when the similar character determination unit determines that the character is a similar character, the registered seal image data and the read seal image data are determined not to match.