KR101650865B1 - Method and apparatus for comparing check by using image process - Google Patents

Abstract

The present invention relates to a method of comparing checks by image processing, comprising the steps of acquiring a scan image for a check, extracting a region of interest (ROI) from a reference scan image corresponding to the scan image and check, , Subdividing each ROI into a plurality of sub-ROIs of a relatively smaller size than the ROI, aligning each ROI based on pixel information of the plurality of ROIs, And comparing the pixel information of the region.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of comparing a check using image processing,

The present invention relates to a method and apparatus for comparing checks by image processing, and more particularly, to a method and apparatus for comparing checks by image processing, and more particularly, And provides a comparison result based on a scanned image for a check, and a method and apparatus for comparing a check through image processing.

Checks are securities in which the issuer entrusts a payment (financial institution) with a certain amount of payment to the recipient or other legitimate holder.

A check shall be made on the basis of the designated method of the addressee, the registered check in which only the name of the addressee is filled in, the indicative check indicating the payment of the person's name to the person other than the name of the addressee, A cashier's check, a cashier's check, a cashier's check, a cashier's check, a cashier's check, a cashier's check, a cashier's cashier's check, The name of the recipient and an undelivered check that does not contain any information to be paid to the bearer.

In addition, checks may be issued in accordance with the issuance format, such as a check, which is issued by the issuer within the scope of check funds in the bank, which has entered into a checking account contract with the bank, a self-check check by the issuer himself as the payee, A cashier's check, a fiduciary check provided by a third party to provide the funds to be paid, and a check on the date of the issuance of the date of issuance with a future date.

In this way, the checks are divided into various types, but in general, the recipient must present the check to the financial institution in order to receive the check issued by the financial institution in money.

The financial institution confirms whether the check is forged or altered using the information stored at the time of issuance, and if there is no abnormality, the transaction is made in such a way that the amount stated in the check is paid to the user.

Prior art relating to the present invention is disclosed in Korean Patent Laid-Open Publication No. 10-2005-0057825 (published on June 16, 2005, entitled "Inventory processing system and method of processing check service using the same).

In the case of a check transaction through an existing method, a financial institution has to check whether a check requested for cash payment by a person is falsified or altered. Therefore, forgery or alteration is detected according to the ability of the person performing the job There has been a problem that there is a difference in accuracy, and there is a problem in that, in the case of a check forgery or falsification by the gathering of employees or conspirators of a financial institution, it can not be prevented originally.

In addition, there is a problem that it takes a long time because a person has to collate the information of the check stored in advance at the time of issuing the check and the check requested to be paid.

Therefore, there is a growing need for technology that allows the transaction to proceed faster and more accurately when paying cash for checks.

The present invention has been made in view of the above-mentioned needs, and it is an object of the present invention to provide a system and method for automatically checking whether a check is the same as a check at the time of issuance, And provides a comparison result based on a scanned image for a check, and a method and apparatus for comparing a check through image processing.

According to an aspect of the present invention, there is provided a method of comparing a check through image processing, comprising: obtaining a scan image for a check; Extracting a region of interest (ROI) from the scan image and the reference scan image corresponding to the check, respectively; Subdividing each of the regions of interest into a plurality of sub-ROIs of a relatively smaller size than the ROI; Aligning the respective regions of interest based on pixel information of the plurality of sub-regions of interest; And comparing pixel information of each of the aligned ROIs.

The present invention further includes converting the scan image and the reference scan image into a gray scale image before extracting the region of interest from the scan image and the reference scan image.

The method may further include determining whether the scan image and the reference scan image are comparable after acquiring the scan image for the check, wherein when the comparable scan image is acquired, And extracts the ROIs from the reference scan image.

In the present invention, the step of aligning each of the ROIs may include: converting each pixel information included in the plurality of ROIs into a bit string of a bit format; Determining a degree of similarity of the converted bit stream for each of the plurality of sub-areas of interest; And aligning the respective ROIs based on positions of sub-ROIs having the highest degree of similarity of the bitstreams.

In aligning the respective regions of interest of the present invention, the pixel information includes a gray scale value of the pixel and the bitstream is determined based on the magnitude of the gray scale value of each pixel .

The method may further include calculating and outputting a bit stream difference of pixel information included in each of the sorted ROIs after comparing the pixel information of each aligned ROI, Highlighting the sub-area of interest and outputting the highlighted sub-area of interest.

The step of calculating and outputting the bit stream difference of the pixel information included in each of the aligned ROIs may further include the step of converting the ROI into a plurality of second ROIs Subdividing; And calculating a difference of bit strings obtained by converting each pixel information included in the plurality of second sub-interest areas, wherein the plurality of second sub-interest areas are subdivided to overlap each other do.

The present invention is characterized by further comprising the step of normalizing each of the sorted ROIs before comparing pixel information of the aligned ROIs.

An apparatus for comparing a check through an image processing according to an aspect of the present invention includes a region of interest extractor for extracting a region of interest (ROI) from a scan image for a check and a reference scan image corresponding to the check; A sub-ROI extractor for subdividing the ROI into a plurality of sub-ROIs of a size relatively smaller than the ROI; An alignment unit arranged to align the respective regions of interest based on pixel information of the plurality of sub-ROIs; And a controller for comparing pixel information of each of the aligned ROIs.

The present invention is further characterized by a memory unit storing a reference scan image corresponding to the check.

The present invention further includes an image converting unit for converting the scan image and the reference scan image into a gray scale image, respectively.

The present invention further includes an input unit for obtaining at least one of the scan image and the reference scan image, and the input unit further determines whether the scan image and the reference scan image are comparable.

In the present invention, the aligning unit converts each pixel information included in the plurality of sub-interest areas into a bit string in a bit format, and determines the similarities of the converted bit strings to the plurality of sub-ROIs And aligning the respective ROIs based on positions of sub-ROIs having the highest degree of similarity of the bitstreams.

In the present invention, the pixel information includes a gray scale value of the pixel, and the alignment unit determines the bit string based on the magnitude of the gray scale value of each pixel.

The present invention further includes an output unit for outputting a result of comparing the pixel information of each of the aligned ROIs based on the control of the controller, wherein when the bit stream difference of the pixel information is equal to or greater than a preset reference value Highlighting the sub-region of interest and outputting it.

In the present invention, the controller may subdivide the sub-region of interest into a plurality of second sub-regions of interest that are relatively smaller than the sub-region of interest through the sub-region of interest extractor, And comparing the bit string difference obtained by converting each pixel information, wherein the plurality of second sub-ROIs are subdivided to overlap each other.

In the present invention, the control unit may normalize each of the ROIs aligned through the ROI, and compare pixel information of each normalized ROI.

According to the present invention, a financial institution can quickly and accurately present a scanned image at the time of issuing a check and a comparison result of a scanned image at the time of paying cash for the check, thereby prompting the user to easily forgive or alter the check So that it can be confirmed.

In addition, the present invention converts each scan image into a grayscale image and compares the results of the normalization to determine the color density, resolution, and color density that can occur when a check image is acquired by different types of image acquisition devices Brightness and the like can be reduced.

1 is a block diagram of an apparatus for comparing checks by image processing according to an embodiment of the present invention.
FIG. 2 is a diagram showing an example of an image which can not be compared.
FIG. 3 is a diagram illustrating an area of interest of a check set in an apparatus for comparing checks by image processing according to an exemplary embodiment of the present invention.
4 is a view showing a result of a comparison between a scan image and a reference scan image in an apparatus for comparing checks by image processing according to an embodiment of the present invention.
FIG. 5 is a flowchart illustrating a method of comparing a check through image processing according to an exemplary embodiment of the present invention. Referring to FIG.

Hereinafter, a method of comparing checks by image processing according to an embodiment of the present invention and its apparatus will be described in detail with reference to the accompanying drawings. In this process, the thicknesses of the lines and the sizes of the components shown in the drawings may be exaggerated for clarity and convenience of explanation. In addition, the terms described below are defined in consideration of the functions of the present invention, which may vary depending on the intention or custom of the user, the operator. Therefore, definitions of these terms should be made based on the contents throughout this specification.

1 is a block diagram of an apparatus for comparing checks by image processing according to an embodiment of the present invention.

1, the apparatus for comparing checks through image processing includes an input unit 10, an image converting unit 20, a ROI extracting unit 40, a ROI extracting unit 50, an aligning unit 60, (70).

The input unit 10 is a configuration for acquiring a scan image for a check. As the input unit 10, various media processing apparatuses having a function of obtaining images of media such as banknotes, checks, and securities can be employed.

The input unit 10 acquires a scanned image for the check to compare a check to be paid with cash to a predetermined reference scan image. Specifically, the input unit 10 includes a device such as a scanner (not shown) Or receive a scan image from a separate scan image acquisition unit (not shown).

At this time, the reference scan image means an image scanned in advance at the time of issuing a specific check to a scan image corresponding to a check to be paid in cash. That is, in this embodiment, an image obtained by scanning a check when a cash payment is requested is referred to as a scan image, and an image obtained by scanning a check at the time of issuance is referred to as a reference scan image.

In this embodiment, in order to provide a criterion for judging whether a check requested to be paid by a user has been altered, corrupted, or damaged, a check is made to the cash payment request and a reference scan Compare images.

Also, in the present embodiment, the input unit 10 determines whether the acquired scan image and reference scan image are comparable. In other words, a check which is too damaged is not normally compared even if it is scanned, and even when a normal check is scanned, it is impossible to compare the check if it is scanned with a lot of errors.

FIG. 2 is a diagram showing an example of an image which can not be compared.

As shown in FIG. 2, when a check is scanned in an oblique state, areas of a check to be compared do not coincide with each other, so that the result of comparison with a reference scan image can not be relied upon.

Therefore, if the slope of the check area included in the scanned image exceeds the reference value or the check is damaged beyond the predetermined area, the input unit 10 determines that comparison is impossible through the check.

The image converting unit 20 converts the scan image and the reference scan image acquired through the input unit 10 into a gray scale image.

As described above, in this embodiment, a scan image for a check requested to be paid by a user is compared with a reference scan image for a previously stored check, and a scanned image is scanned with a different scanner May be obtained.

In this case, as different scanners use different sensors, optical characteristics such as density, brightness, and resolution of colors included in the scanned image may be different.

Therefore, in this embodiment, by converting the scan image and the reference scan image into the grayscale image, it is possible to reduce the influence due to differences in color density, brightness, and resolution due to the use of different scanners.

The memory unit 30 stores a reference scan image obtained by scanning the check at the time of the check issuance. Here, the memory unit 30 may be constituted by a memory device, which is an independent unit, or may be constituted by various servers such as a web-based server, a database, or a cloud-based data storage unit, Means to include any form of storage.

The ROI extracting unit 40 extracts ROIs from the scan image and the reference scan image.

That is, comparing all the areas of the check may take a long time to calculate the result because the calculation amount is too large. Therefore, in the present embodiment, only the region of interest previously determined for the scan image and the reference scan image is extracted and compared.

Accordingly, the area of interest refers to an area including information for judging whether a particular check is struck, modulated or damaged. The area of interest for a specific check may be set differently according to the bank.

For example, the area of interest may be set to include MICR (Magnetic Ink Character Recognition) printed in magnetic ink, the serial number at the top of the check, the amount, the logo, the date of check issuance,

FIG. 3 is a diagram illustrating an area of interest of a check set in an apparatus for comparing checks by image processing according to an exemplary embodiment of the present invention.

3 shows the serial number, (b) the bank logo, (c) the amount, (d) the check issuance date, (e) the responsible person seal, (f) the MICR, As described above, since such an area of interest is only one example, it is possible to set it to include more various areas.

As described above, in the present embodiment, the ROI extracting unit 40 extracts a ROI that is preset in the scan image and the reference scan image for effective comparison between the scan image and the reference scan image.

The sub-ROI extractor 50 subdivides the ROI extracted by the ROI extractor 40 into a plurality of sub-ROIs having a relatively smaller size than the ROI.

As described above, in the present embodiment, the scan image and the reference scan image are compared with each other based on a more detailed sub-area of interest instead of comparing the scan image and the reference scan image based on a wide range of interest regions, A more accurate comparison result can be obtained.

That is, the interest area such as the bank logo or the responsible person's seal shown in FIG. 3 is formed in a relatively narrow range as compared with other interest areas, but since the interest areas such as the money amount and the MICR are formed in a wide range, In the case of comparison, the difference in the minor area can be ignored, and an error may occur in the comparison result.

Therefore, in this embodiment, the reliability of the comparison result can be improved by extracting the sub-ROIs that further subdivide the ROIs.

In addition, in the present embodiment, in order to align the regions of interest to be compared with each other by comparing the detailed region of interest extracted by subdividing the region of interest extracted from the scan image and the reference scan image through the arranging unit 60, Subdividing the region of interest into sub-regions of interest.

The alignment unit 60 aligns each of the ROIs based on the respective pixel information included in the plurality of sub-ROIs. Specifically, the aligning unit 60 converts each pixel information included in a plurality of sub-interest areas into a bit string of a bit format, and determines the similarity of the converted bit string to a plurality of sub- It can be determined which sub-region of interest is most similar in each of the scan image and the reference scan image.

When the check is waved or scrolled in a collapsed state when acquiring a scan image for a check, the positions of the respective regions of interest extracted from the scan image and the reference scan image are minutely different.

In this way, since the range of coordinates occupied by the region of interest in the scanned image and the range of coordinates occupied by the region of interest in the reference scan image can be changed, when comparing each of the regions of interest or sub-region of interest, There is a lot of problems.

Therefore, in the present embodiment, each pixel information included in the sub-ROI extracted by the ROI extracting unit 50 is converted into a bit stream, and based on the similarity of each bit stream, .

That is, the aligning unit 60 arranges the scanned image and the region of interest of the reference scan image on the basis of the position coordinates of each sub-area of interest in which bit-stream similarity is most similar.

More specifically, the aligning unit 60 converts information on a gray scale value of a pixel into a bit string. A reference pixel is determined from among pixels belonging to a specific range, and a gray scale value The information on all the pixels included in the sub-region of interest may be converted into a bit stream based on the magnitude relationship of the sub-region of interest.

For example, based on the magnitude relation between the reference pixel and the gray scale value of the remaining eight pixels, the bit information of the pixel larger than the gray scale value of the reference pixel is set to 1, The pixel information can be simply converted into the bit stream in such a manner that the bit information of the pixel smaller than the gray scale value of the pixel is set to zero.

At this time, when nine pixels are converted into bit streams, all the pixel information included in the sub-region of interest is converted into bit streams by sequentially converting another nine pixels included in the sub-region of interest into bit streams .

The alignment unit 60 aligns the region of interest based on the similarity between the bit stream corresponding to the sub-region of interest of the scan image and the bit stream corresponding to the sub-region of interest of the reference scan image.

In particular, since the ROI is subdivided into a plurality of sub-ROIs, the aligner 60 acquires a plurality of bit streams corresponding to a plurality of sub-ROIs, You can judge how similar the rows are.

For example, if it is assumed that the ROI is subdivided into n sub-ROIs, the aligner 60 obtains n bitstreams for the ROI of the scan image, thereby obtaining n bit strings.

Accordingly, the aligner 60 can compare the n bit strings of the sub-ROI of the scan image with the n bit strings of the ROI of the reference scan image to determine which bit strings are most similar to each other .

At this time, it is possible to judge whether the bit strings are similar based on the number of matched bits among the bits included in the bit string. That is, the more similar bits are, the higher the degree of similarity of bit strings is.

As a result of the determination, when the most matching pair of bit strings is found, the aligning unit 60 can know the position coordinates occupied by the corresponding bit string in the scan image and the position coordinates occupied by the corresponding bit string in the reference scan image, The region of interest may be aligned so that the range of interest of the scan image matches the range of interest of the reference scan image.

As described above, in the present embodiment, before the scan image and the reference scan image are compared, it is possible to compare the corresponding regions by matching the regions of interest of the respective images to be compared, thereby improving the reliability of the comparison result.

The controller 70 compares the pixel information included in the ROI of the aligned reference scan image with the ROI of the aligned scan image.

That is, in this embodiment, since the positions of the ROIs to be compared are matched through the arranging unit 60, the controller 70 determines whether the bits at the same position coincide with each other, It is possible to calculate to what extent the bit string is different.

In particular, in this embodiment, the control unit 70 may control the output unit 80 described later so that the bit stream difference is greater than or equal to a preset reference value and the sub-region of interest is highlighted and output.

In other words, a large difference in the bit stream converted from the pixel information included in the specific sub-area of interest indicates that the sub-area of interest may be falsified or altered. Therefore, the control unit 70 may highlight the sub- It can be confirmed that there is a large difference from the check at the time of issuance at the position.

Specifically, when determining whether the bit string difference is equal to or greater than a preset reference value, the controller 70 determines whether an error rate at a point of time when the False Acceptance Rate (FAR) and the False Rejection Rate (FRR) The reference value can be determined by reflecting the Equal Error Rate (ERR).

That is, the error probability can be minimized by determining the reference value through the same error rate at the time when the error rate judged to be normal even though it has been stolen, altered or damaged, and the error rate at which the normal banknote is judged to be distorted, modulated or damaged.

However, since the reference value can be set in various ways, it is not limited to the above-described method.

In addition, in this embodiment, the controller 70 may normalize each aligned region of interest before comparing the pixel information of the region of interest of the scan image and the reference scan image.

Normalization refers to changing a specific object to a regular state according to a certain rule or criterion. As described above, when the same check is scanned by a different scanner at the time of issuing the check and at the time of requesting the cash payment, As well as whether the brightness and boundary are emphasized can be different.

Therefore, in this embodiment, the control unit 70 determines the similarity of each ROI more accurately by normalizing each ROI and comparing the normalized pixel information. The pixel information at this time can also be set to a gray scale value.

To be more specific, the controller 70 normalizes the pixel information of each RO using the z-score method as shown in Equation 1 below, and converts each pixel information into values of 0 to 255 .

(이 때, g(x,y)는 각 픽셀의 z-score 값, max와 min은 각 픽셀의 z-score 값의 최대값과 최소값을 의미함)

(Where g (x, y) is the z-score value of each pixel, and max and min are the maximum and minimum values of the z-score value of each pixel)

In this way, the scanned image of the check scanned by the different scanners and the reference scan image are normalized, thereby reducing the difference in the scanned images caused by the different scans of the brightness during the scanning process.

However, the present embodiment is not limited to this, and it is also possible to perform normalization using various other known or unknown methods other than the normalization method described above.

In addition, in this embodiment, the control unit 70 sub-sub-populates the sub-region of interest into a plurality of second sub-regions of interest that are relatively smaller in size than the sub-region of interest through the sub-region of interest extraction unit 50, The second sub-region of interest is subdivided to overlap each other.

Then, the controller 70 compares the scan image and the reference scan image based on the pixel information included in the overlapped second sub-area of interest.

That is, the controller 70 compares the pixel information included in the sub-region of interest with subdivided regions of interest and determines how much the scan image differs from the reference scan image based on the ratio of the number of pixels with respect to the total number of pixels. In the case where a small area is damaged or staggered or modulated compared to the entire sub-area of interest, it is judged that there is no problem in the scan image because the area is small irrespective of the importance of the area.

Accordingly, in this embodiment, the second sub-region of interest is subdivided so that the second sub-region of interest overlaps with the sub-region of interest.

In other words, when the boundary between the sub-interest areas is staggered, the modulation or the degradation may not be detected, the second sub-areas of interest may be overlapped with each other, So that the pixel information can be compared with respect to the boundary.

The output unit 80 outputs the result of the comparison of the pixel information of the region of interest of the scan image and the reference scan image under the control of the control unit 70.

In particular, in the present embodiment, the output unit 80 compares the pixel information included in each ROI, and when the bit string difference is equal to or greater than a predetermined reference value, the ROI unit highlights the ROI, Modulated, or tampered with.

4 is a view showing a result of a comparison between a scan image and a reference scan image in an apparatus for comparing checks by image processing according to an embodiment of the present invention.

4A is a view showing a case where the output unit 80 highlights a specific sub-region of interest based on a result of comparing the region of interest of the scan image and the reference scan image, (C) is a diagram showing a region of interest of a scanned image.

4 (B) and 4 (C), an amount of 5 was scanned in the region of interest in (B), and 0 was scanned in the region of interest in (C) have. Therefore, in this embodiment, the output unit 80 highlights the sub-interest area having a difference as shown in (A), and outputs the highlighted sub-area of interest, so that the difference between the corresponding positions can be confirmed. The method of highlighting and outputting is a method of outputting colors with different colors, such as a method of displaying the image in the form of a rectangle, a circle, or an arrow, as shown in (A) And a method of enlarging and outputting the image data. In addition, the degree of difference between the scan image and the reference scan image may be output in such a manner as scoring, scaling, tabulating, and graphing.

FIG. 5 is a flowchart illustrating a method of comparing a check through image processing according to an exemplary embodiment of the present invention. Referring to FIG.

5, an input unit 10 acquires a scan image for a check (S10), and acquires a scan image and a memory (S20) whether the reference scan image stored in the storage unit 30 is a comparable image.

In other words, a check that is too damaged may not be normally compared even if it is scanned, and even when a normal check is scanned, it is impossible to compare the check if it is scanned in a much wrong state. Before comparison, it is first determined whether each scan image is in a state where comparison is possible.

If it is determined that the comparison is possible, the image converting unit 20 converts the scan image and the reference scan image into a gray scale image (S30).

The scanned image of the check requesting the cash payment from the user and the reference scan image of the stored check may be acquired by different scanners. In this case, different scanners may be included in the scanned image by using different sensors The density, brightness, resolution, and the like of the colors may be different.

Therefore, in this embodiment, by converting the scan image and the reference scan image into the grayscale image, it is possible to reduce the influence of the color density or resolution due to the use of different scanners.

Then, the ROI extractor 40 extracts the ROIs of the predetermined positions in the scan image and the reference scan image, respectively (S40).

That is, comparing all the areas of the check may take a long time to calculate the result because the calculation amount is too large. Therefore, in the present embodiment, only the region of interest previously determined for the scan image and the reference scan image is extracted and compared.

Accordingly, the area of interest refers to an area including information for judging whether a particular check is struck, modulated or damaged. The area of interest for a specific check may be set differently according to the bank.

Then, the sub-ROI extracting unit 50 subdivides each ROI into a plurality of sub-ROIs of a size that is relatively smaller than the ROI (S50).

That is, instead of comparing the scan image with the reference scan image on the basis of a wide range of interest areas, the scan image and the reference scan image are compared with each other based on a subdivided interest area, Accurate comparison results can be obtained.

In addition, in the present embodiment, the interest region is subdivided into sub-interest regions in order to align the positions of the ROIs to be compared through comparison of the ROIs.

Then, the aligning unit 60 aligns the respective regions of interest based on the pixel information of the plurality of sub-regions of interest (S60).

Specifically, in this embodiment, the region of interest is aligned based on the pixel information including the gray scale value of the pixel. The process of aligning the region of interest will be described below.

The sorting unit 60 converts each pixel information included in the plurality of sub-interest areas into a bit string of bit format (S62), determines the similarity of the converted bit string to the entire ROI (S64) The respective regions of interest are aligned based on the positions of the sub-ROIs having the highest degree of similarity of the columns (S66).

When the check is waved or scrolled in a collapsed state when acquiring a scan image for a check, the positions of the respective regions of interest extracted from the scan image and the reference scan image are minutely different.

That is, since the range of the coordinates occupied by the region of interest in the scanned image and the range of the coordinates occupied by the region of interest in the reference scan image can be changed, when the respective regions of interest or sub-region of interest are compared, There is a lot to happen.

Therefore, in the present embodiment, each pixel information included in the sub-ROI extracted by the ROI extracting unit 50 is converted into a bit stream, and based on the similarity of each bit stream, .

Next, the control unit 70 normalizes each of the ROIs sorted in step S60, compares the normalized ROIs, and outputs the comparison result through the output unit 80 (S80 ).

As described above, this occurs when scanning the same check at the time of issuing the check and at the time of requesting the cash payment to different scanners. In this case, not only the color density but also the brightness and boundary emphasis may be different have.

Therefore, in this embodiment, the control unit 70 determines the similarity of each ROI more accurately by normalizing each ROI and comparing the normalized pixel information.

According to this embodiment, a financial institution can quickly and accurately present a scanned image at the time of issuing a check and a comparison result of the scanned image at the time of payment of cash for the check, thereby prompting the user to easily forgive or counterfeit the check .

In addition, in the present embodiment, each scan image is converted into a grayscale image and compared with the normalized result, it is possible to reduce the influence of color density, brightness, resolution, and the like which may occur when a check is scanned by different scanners have.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is clearly understood that the same is by way of illustration and example only and is not to be taken by way of limitation, I will understand. Accordingly, the true scope of the present invention should be determined by the following claims.

10: Input unit
20:
30: memory unit
40: ROI extracting unit
50: Sub-interest region extracting unit
60:
70:
80: Output section

Claims (17)

Obtaining a scanned image for a check;
Extracting a region of interest (ROI) from the scan image and the reference scan image corresponding to the check, respectively;
Subdividing each of the regions of interest into a plurality of sub-ROIs of a relatively smaller size than the ROI;
Aligning the respective regions of interest based on pixel information of the plurality of sub-regions of interest; And
Comparing pixel information of each of the aligned ROIs
A method of comparing a check through an image processing including the image processing.
The method according to claim 1,
And converting the scan image and the reference scan image into a gray scale image before extracting the ROI from the scan image and the reference scan image, Way.
The method according to claim 1,
Determining whether the scan image and the reference scan image are comparable images after acquiring the scan image for the check, wherein when the comparable scan image is acquired, the scan image and the reference scan image Wherein the region of interest is extracted from the region of interest.
The method according to claim 1,
Wherein the step of aligning each of the ROIs includes: converting each pixel information included in the plurality of sub-ROIs into a bitstream of a Bit format;
Determining a degree of similarity of the converted bit stream for each of the plurality of sub-areas of interest; And
Aligning the respective ROIs based on positions of sub-ROIs having the highest degree of similarity of the bitstreams
And comparing the checks with image processing.
5. The method of claim 4,
Characterized in that in the step of aligning each of the areas of interest, the pixel information comprises a gray scale value of the pixel and the bit stream is determined based on the magnitude of the gray scale value of each pixel A comparison method of check through image processing.
6. The method of claim 5,
Calculating and outputting a bit stream difference of pixel information included in each of the ordered ROIs after comparing the pixel information of each aligned ROI, Highlighting the sub-area of interest and outputting the highlighted sub-area of interest.
The method according to claim 6,
The step of calculating and outputting the bit stream difference of the pixel information included in each of the sorted ROIs may include subdividing the ROI into a plurality of second ROIs of a relatively smaller size than the ROI ; And
And calculating a bit string difference by converting each pixel information included in the plurality of second sub-ROIs, wherein the plurality of second sub-ROIs are subdivided to overlap each other A comparison method of check through image processing.
The method according to claim 1,
Further comprising the step of normalizing each of the sorted ROIs before comparing the pixel information of the aligned ROIs.
An ROI extracting unit that extracts a region of interest (ROI) from a scan image for a check and a reference scan image corresponding to the check;
A sub-ROI extractor for subdividing the ROI into a plurality of sub-ROIs of a size relatively smaller than the ROI;
An alignment unit arranged to align the respective regions of interest based on pixel information of the plurality of sub-ROIs; And
A controller for comparing pixel information of each of the aligned ROIs,
And a control unit for comparing the check through the image processing.
10. The method of claim 9,
Further comprising a memory unit for storing a reference scan image corresponding to the check.
10. The method of claim 9,
Further comprising an image converting unit for converting the scan image and the reference scan image into a gray scale image, respectively.
10. The method of claim 9,
And an input unit for acquiring at least one of the scan image and the reference scan image,
Wherein the input unit further determines whether the scan image and the reference scan image are comparable images.
10. The method of claim 9,
The arrangement unit converts each pixel information included in the plurality of sub-interest areas into a bit string in a bit format, determines similarities between the converted bit strings for the plurality of sub-ROIs, And aligning the respective regions of interest based on the positions of the sub-ROIs having the highest degree of similarity of the columns.
14. The method of claim 13,
Wherein the pixel information includes a gray scale value of the pixel and the sorting unit determines the bit string based on a magnitude of a gray scale value of each pixel. Device.
15. The method of claim 14,
Further comprising an output unit for outputting a result of comparing pixel information of each of the sorted ROIs based on the control of the controller, wherein if the bitstream difference of the pixel information is equal to or greater than a preset reference value, Area is highlighted and output. The device for comparing checks by image processing.
16. The method of claim 15,
Wherein the controller sub-divides the sub-region of interest into a plurality of second sub-regions of interest that are relatively smaller in size than the sub-region of interest through the sub-region of interest extractor, and wherein each pixel included in the plurality of second sub- Wherein the second sub-interest areas are subdivided so as to overlap with each other, wherein the plurality of second sub-areas of interest are subdivided so as to overlap each other.
10. The method of claim 9,
Wherein the control unit normalizes each of the ROIs aligned through the ROI and compares pixel information of each of the ROIs with the ROI.

Images