KR101346328B1 - Identification distincting apparatus using degraded images and distincting apparatus using the same - Google Patents

Abstract

Disclosed are an apparatus for distinguishing an identification having a low quality image and a method for distinguishing an identification using the same. An apparatus for distinguishing an identification having a low quality image comprises: a first processing module (110) for receiving identification image data received from the outside, preprocessing the identification image data, and outputting a first result value (1st) corresponding to the identification image data; a second processing module (120) for setting the size of the identification image data, which is preprocessed as the first result value (1st), at a predetermined ratio, and outputting a second result value (2st) corresponding to the set size; a third processing module (130) for receiving the second result value (2st), grouping pixels of multiple areas within the second result value (2st) through a K-means clustering algorithm, and outputting a third result value (3st) corresponding to the grouped pixels; a fourth processing module (140) for setting an area of the grouped pixels, which is the third result value (3st), and outputting a fourth result value (4st) corresponding to the set area; a fifth processing module (150) for outputting a fifth result value (5st) extracting an identification image arranged based on the set area, which is the fourth result value (4st); a sixth processing module (160) for receiving the fifth result value (5st), distinguishing an image type of the fifth result value (5st) using template matching, and outputting a sixth result value (6st) corresponding to the image type; and a seventh processing module (170) for receiving the sixth result value (6st) and extracting a personal information area using a relative location about the matching result. [Reference numerals] (110) First processing module;(120) Second processing module;(130) Third processing module;(140) Fourth processing module;(150) Fifth processing module;(160) Sixth processing module;(170) Seventh processing module;(AA) Image data;(BB) Personal information area extraction

Description

Identification identification apparatus using low quality image and identification method using same {identification distincting apparatus using Degraded Images and distincting apparatus using the same}

The present invention relates to a region of interest detection using an image recognition method, and more particularly, to an identification card having a low quality image and a discrimination method using the same.

As is well known, the identification means using personal identification, user authentication, or personal information protection, for example, national ID card, driver's license, student's card, etc., are not used by individuals. In this case, since the identity is difficult to verify, identification is not easy, and sometimes, even when the owner of the identification means and the actual owner are not the same, it may occur as the identity.

In order to compensate for this problem, for example, development of biometric technologies such as fingerprint recognition, iris recognition, and face recognition is underway. It is attracting attention in various applications because of its low coercivity and less reluctance in a non-contact manner.

Here, human detection technology including face recognition, which is one of biometric technologies, has been studied for a long time as a core technology of biometrics, and has recently been used as a detection technology for biometrics. Along with the expansion of the market related to the market, various developments are being made.

In such an environment, when a camera is mounted on a digital device or a mobile device and a person recognition technology including face recognition is applied, the value added of the product and the increase in sales are expected. For example, a mobile device such as a mobile phone may provide a function of detecting a location of a person, recognizing a human face through image processing, and then changing a facial expression. In this case, it can provide a function of focusing on a person's location, and can be utilized by combining with various technologies.

On the other hand, human detection algorithms including face recognition have been developed in the form of algorithms mainly operating in PC-based environment. When this technology is directly applied to embedded systems, it is impossible to detect people in real time because of relatively insufficient resources and performance. Efficient human detection was difficult due to low or low detection rate.

However, due to the expansion of the service robot field, the market for home robots, such as cleaning robots and entertainment toy robots, is growing in daily life, and human biometric information can be used in portable devices such as mobile phones and digital cameras. As the number of applications increases, the need for high performance real-time human detection technology in embedded systems is growing.

However, the image recognition method used in such an embedded system is limited to a high quality image and is not used in a low quality image.

For example, in the case of ID card, it is not difficult to extract by using the advantages of color transmission and border preservation when transmitting in high quality, but low quality video does not preserve the border and uses binary image, limitations due to lighting, etc. There was a lot of difficulty in recognition.

Disclosure of the Invention An object of the present invention is to provide an identification card discrimination apparatus having a low quality image and a method of identifying the same having a low quality image that can extract a personal information region even in a low quality image.

An identification card having a low quality image according to an embodiment of the present invention for solving the above problems is a first processing for receiving the identification image data received from the outside and pre-processing to output a corresponding first result value (1st) Module 110; A second processing module 120 for setting the size of the pre-processed identification image data that is the first result value 1st at a preset ratio and outputting a second result value 2st corresponding thereto; A third result of receiving the second result value 2st, pixel grouping a plurality of regions in the second result value 2st through a K-means clustering algorithm, and outputting a third result value 3st corresponding thereto; Processing module 130; A fourth processing module 140 for setting an area of the grouped pixels that is the third result value 3st and outputting a fourth result value 4st corresponding thereto; A fifth processing module 150 for outputting a fifth result value 5st obtained by extracting an ID image arranged based on the set area that is the fourth result value 4st; After receiving the fifth result value 5st, determining an image type of the fifth result value 5st using template matching and outputting a sixth result value 6st corresponding thereto; 6 processing module 160; And a seventh processing module 170 which receives the sixth result value 6st and extracts a personal information area using a relative position of the matching result.

According to an embodiment of the present invention for solving the above problems, a low-quality identification card identification method using the identification card identification device of claim 1 includes an image preprocessing step of pre-processing identification image data received from the outside; A ratio setting step of setting the size of the pre-processed identification image data to a preset ratio (S120); A pixel grouping step (S130) of pixel grouping a plurality of areas of the image data using a K-means clustering algorithm after the ratio setting step (S120); an area setting step (S140) of setting an area of grouped pixels; An image extraction step (S150) of extracting an ID image arranged based on the set area; A determination step (S160) of determining the image type using template matching; An extraction step S170 of extracting the personal information area using the relative position of the matching result is included.

The ratio setting step S120 may include: a first step S121 of converting the ID card image data into a black and white image and then reducing the ratio of the width and width of the image to a size of 30% respectively; A second step (S122) of performing a convolution operation on the reduced image with a median filter having a size of 3x3; And a third step (S123) of reducing the image again to 50% of the width and width, respectively.

The second step (S122) is characterized in that the step performed to improve the processing speed.

The second step S122 may further include binarizing the identification card image.

The pixel grouping step (S130) may include a first step (S131) of setting an initial value of the K; A second step (S132) of performing the K-means clustering algorithm; A third step (S133) of obtaining the midpoints of all the pixels of the grouped clusters and comparing the distances between the midpoints and the identification image size to confirm the validity of K; And if the K value is not valid, repeating the second step (S132) to the third step (S133) after reducing the K by one.

The area setting step (S140) may include: a storage step (S141) of storing four coordinate positions corresponding to the minimum and maximum values of the X-axis, the minimum and maximum values of the Y-axis, and the coordinate plane which is the area of the grouped pixels; After the storing step, determining a validity of vertices of the remaining pixels by measuring a distance between the pixel of the minimum value of the X axis and the remaining pixels except the X axis (S142); And a reset step (S143) of resetting the vertex by using the minimum and maximum values of the storing step when the vertices of the remaining pixels are not valid.

The resetting step (S143) is a step of setting the minimum values of the X-axis and the Y-axis as vertices at the upper left, and the maximum values of the X-axis and the Y-axis as vertices at the lower right.

The image extracting step (S150) may include a first step (S151) of extracting auxiliary images of four regions corresponding to the minimum and maximum values of the X and Y axes; Setting a center point of the auxiliary image (S152); A third step (S153) of calculating an angle for sub-image alignment of the four areas; A fourth step (S154) of rotating the auxiliary image by the arrangement angle calculated in the third step; A fifth step (S155) of calculating a black pixel of the X-axis and the Y-axis histogram after rotating the rotated image from -10 ° to 10 ° in the fourth step (S154); And a sixth step S156 of obtaining an aligned image by rotating the image by applying the rotation angle having the largest number calculated in the fifth step S155.

The determining step S160 may include performing template matching within a preset area range from an upper left vertex in the aligned image, and the performing range of the template matching may include: Characterized in that it is up to the position including the template image from the upper left.

The template matching uses NCC (Normalized Cross Correlation), and is performed in units of two pixels to shorten execution time.

The image extracting step S150 may further include performing the template matching by rotating the aligned image 180 degrees.

In the template matching, when the maximum value of the NCC is less than 0.7, template matching is performed using a template of another ID image.

According to the present invention, there is an advantage of quickly recognizing and processing through a process of performing a process of recognizing and extracting the image by reducing the initial stage and omitting pixel matching in a range that does not significantly affect performance in template matching.

In addition, since the sensitive personal information region can be extracted using the relative position after template matching, it can be applied in the process of extracting or hiding the personal information during network transmission or by extracting the information due to the need for personal information.

1 is a block diagram showing an identification card having a low quality image according to an embodiment of the present invention.
FIG. 2 is a flowchart for explaining an identification card having a low quality image using the apparatus of FIG. 1.
FIG. 3 is a flowchart illustrating the ratio setting step S120 shown in FIG. 2 in more detail.
4 is an exemplary view showing a series of processes occurring in the image preprocessing process, (a) shows the image reduced and the black and white image of the identification card, (b) shows (a) after the median filtering, (c) is an exemplary diagram showing image reduction and binarized identification.
FIG. 5 is a flowchart illustrating the pixel grouping step S130 shown in FIG. 2 in more detail.
6A and 6B are exemplary views showing two IDs to be classified through K-means clustering.
FIG. 7 is a flowchart illustrating in detail the region setting step S140 shown in FIG. 2.
8A is an exemplary diagram illustrating vertex validation using Pythagorean theorem.
FIG. 9 is a flowchart illustrating in detail the image extraction step S170 illustrated in FIG. 2.
(A) of FIG. 10 shows a first rotation by a vertex, (b) shows a foreground calculation method by an equal interval inspection line, and (c) is an exemplary view showing an example of a second rotation process.
11 is an exemplary view showing a template made for identification card type recognition and region extraction.
12 is an exemplary view showing an example of the personal information area extracted using the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a detailed description of preferred embodiments of the present invention will be given with reference to the accompanying drawings. In the following description of the present invention, detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

Embodiments in accordance with the concepts of the present invention can make various changes and have various forms, so that specific embodiments are illustrated in the drawings and described in detail in this specification or application. It is to be understood, however, that it is not intended to limit the embodiments according to the concepts of the present invention to the particular forms of disclosure, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between. Other expressions that describe the relationship between components, such as "between" and "between" or "neighboring to" and "directly adjacent to" should be interpreted as well.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, the terms "comprises ",or" having ", or the like, specify that there is a stated feature, number, step, operation, , Steps, operations, components, parts, or combinations thereof, as a matter of principle.

Hereinafter, with reference to the drawings to describe the present invention in more detail.

1 is a block diagram showing an identification card having a low quality image according to an embodiment of the present invention.

FIG. 2 is a flowchart for explaining an identification card having a low quality image using the apparatus of FIG. 1.

FIG. 3 is a flowchart illustrating the ratio setting step S120 shown in FIG. 2 in more detail.

4 is an exemplary view showing a series of processes occurring in the image preprocessing process, (a) shows the image reduced and the black and white image of the identification card, (b) shows (a) after the median filtering, (c) is an exemplary diagram showing image reduction and binarized identification.

FIG. 5 is a flowchart illustrating the pixel grouping step S130 shown in FIG. 2 in more detail.

6A and 6B are exemplary views showing two IDs to be classified through K-means clustering.

FIG. 7 is a flowchart illustrating in detail the region setting step S140 shown in FIG. 2.

8A is an exemplary diagram illustrating vertex validation using Pythagorean theorem.

FIG. 9 is a flowchart illustrating in detail the image extraction step S170 illustrated in FIG. 2.

(A) of FIG. 10 shows a first rotation by a vertex, (b) shows a foreground calculation method by an equal interval inspection line, and (c) is an exemplary view showing an example of a second rotation process.

11 is an exemplary view showing a template made for identification card type recognition and region extraction.

12 is an exemplary view showing an example of the personal information area extracted using the present invention.

As shown in FIG. 1, the determination apparatus 100 of the present invention includes first processing modules 110 and 110 to seventh processing module 170.

The first processing module 110 receives ID image data received from the outside and preprocesses the ID image data to output a first result value 1st corresponding thereto.

The second processing module 120 sets the size of the pre-processed ID image data, which is the first result value 1st, at a preset ratio and outputs a second result value 2st corresponding thereto.

The third processing module 130 receives the second result value 2st, pixel-groups a plurality of regions in the second result value 2st through a K-means clustering algorithm, and corresponds to a third result corresponding thereto. Output the value (3st).

The fourth processing module 140 sets an area of the grouped pixels that is the third result value 3st and outputs a fourth result value 4st corresponding thereto.

The fifth processing module 150 outputs a fifth result value 5st obtained by extracting an ID image arranged based on the set area that is the fourth result value 4st.

After receiving the fifth result value 5st, the sixth processing module determines an image type of the fifth result value 5st by using template matching and corresponds to the sixth result value ( 6st).

The seventh processing module receives the sixth result value 6st and extracts a personal information area using the relative position of the matching result.

As shown in FIG. 2, the identification card identification method S100 using the low quality image of the present invention includes an image preprocessing step S110, a ratio setting step S120, an area setting step S130, a determination step S140, and the like. Extraction step (S150) is included.

The image preprocessing step S110 may be a step of preprocessing ID image data received from the outside.

The ratio setting step S120 may be a step of setting the size of the pre-processed identification image data to a preset ratio.

The area setting step (S140) may include a pixel grouping step (S130) of pixel grouping a plurality of areas of the image data using a K-means clustering algorithm after the ratio setting step (S120); It may be a step of setting an area of the grouped pixels.

The determining step (S160) may include an image extraction step (S150) of extracting an ID image arranged based on the set area; The image type may be determined by using template matching.

The extracting step S170 may be a step of extracting a personal information area using a relative position with respect to the matching result.

More specifically, the ratio setting step S120 includes first to third steps S121 to S123.

The first step S121 may be a step of reducing the ratio of the width and width of the image to a size of 30% after converting the ID card image data into a black and white image.

The second step S122 may be a step of performing a convolution operation on the reduced image with a median filter having a size of 3 × 3.

Here, the second step S122 is performed to improve processing speed, and may include binarizing the identification image.

The third step S123 may be a step of reducing the image again to 50% of the width and width.

As shown in FIG. 5, the pixel grouping step S130 may include first steps S131 to fourth steps S134.

The first step S131 may be a step of setting an initial value of K.

The second step S132 may be a step of performing the K-means clustering algorithm.

The third step S133 may be a step of determining the validity of K by obtaining midpoints of all the pixels of the grouped clusters and comparing distances between the midpoints and an ID image size.

The fourth step (S134) may be a step of repeating the second step (S132) to the third step (S134) after reducing the K by 1 when the K value is not valid. .

As shown in FIG. 7, the area setting step S140 includes a storing step S141, a validity determining step S142, and a resetting step S143.

The storing step S141 may be a step of storing four pixel positions corresponding to the minimum and maximum values of the X-axis, the minimum and maximum values of the Y-axis, and the coordinate plane which is the area of the grouped pixels.

The validity determining step S142 may be a step of determining the validity of the vertices of the remaining pixels by measuring a distance between the pixel of the minimum value of the X-axis and the remaining pixels except the X-axis after the storing step S141.

The resetting step S143 may be a step of resetting vertices using minimum and maximum values of the storing step when the vertices of the remaining pixels are not valid.

The resetting step (S143) may be a step of setting the minimum value of the X-axis and the Y-axis as a vertex on the upper left side, and the maximum value of the X-axis and the Y-axis as the vertex on the lower right side.

As shown in FIG. 9, the image extraction step S150 includes first steps S151 to fifth steps S155.

The first step S151 may be a step of extracting auxiliary images of four areas corresponding to the minimum and maximum values of the X-axis and the Y-axis.

The second step S152 may be a step of setting a center point of the auxiliary image.

The third step S153 may be a step of calculating an angle for aligning the auxiliary images of the four areas.

The fourth step S154 may be a step of rotating the auxiliary image at the arrangement angle calculated in the third step.

The fifth step S155 may be a step of calculating the black pixels of the X-axis and the Y-axis histogram after rotating the image rotated in the fourth step from -10 ° to 10 °.

The sixth step S156 may be a step of obtaining an aligned image by rotating the image by applying the rotation angle having the largest number calculated in the fifth step.

The determining step S160 may include performing template matching within a preset area range from the upper left corner of the aligned image, wherein the template matching range is performed on the left side of the aligned image. The range is set from the top to a position including the template image.

The template matching uses NCC (Normalized Cross Correlation), and is performed in units of two pixels to shorten execution time.

The image extracting step S170 may further include performing the template matching by rotating the aligned image 180 degrees.

In addition, when the maximum value of the NCC is less than 0.7, the template matching performs a template matching with a template of another ID image.

According to the present invention, there is an advantage of quickly recognizing and processing through a process of performing a process of recognizing and extracting the image by reducing the initial stage and omitting pixel matching in a range that does not significantly affect performance in template matching.

In addition, since the sensitive personal information region can be extracted using the relative position after template matching, it can be applied in the process of extracting or hiding the personal information during network transmission or by extracting the information due to the need for personal information.

It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present invention as defined by the appended claims and their equivalents. And such changes are, of course, within the scope of the claims.

100: determination device
110: first processing module
120: second processing module
130: third processing module
140: fourth processing module
150: fifth processing module
160: sixth processing module
170: seventh processing module

Claims (13)

A first processing module 110 which receives ID card image data received from the outside and preprocesses the first image to output a first result value 1st corresponding thereto;
A second processing module 120 for setting the size of the pre-processed identification image data that is the first result value 1st at a preset ratio and outputting a second result value 2st corresponding thereto;
A third result of receiving the second result value 2st, pixel grouping a plurality of regions in the second result value 2st through a K-means clustering algorithm, and outputting a third result value 3st corresponding thereto; Processing module 130;
A fourth processing module 140 for setting an area of the grouped pixels that is the third result value 3st and outputting a fourth result value 4st corresponding thereto;
A fifth processing module 150 for outputting a fifth result value 5st obtained by extracting an ID image arranged based on the set area that is the fourth result value 4st;
After receiving the fifth result value 5st, determining an image type of the fifth result value 5st using template matching and outputting a sixth result value 6st corresponding thereto; 6 processing module 160; And
And a seventh processing module (170) for receiving the sixth result value (6st) and extracting a personal information area using a relative position of the matching result.
In the low quality identification card identification method using the identification card identification device of claim 1,
An image preprocessing step (S110) of preprocessing the ID image data received from the outside;
A ratio setting step of setting the size of the pre-processed identification image data to a preset ratio (S120);
A pixel grouping step (S130) of pixel grouping a plurality of areas of the image data using a K-means clustering algorithm after the ratio setting step (S120);
An area setting step (S140) of setting an area of grouped pixels;
An image extraction step (S150) of extracting an ID image arranged based on the set area;
A determination step (S160) of determining the image type using template matching;
ID identification method having a low quality image comprising an extraction step (S170) of extracting a personal information area using a relative position to the matching result.
3. The method of claim 2,
The ratio setting step (S120),
A first step (S121) of converting the ID card image data into a black and white image and then reducing the ratio of the width and width of the image to a size of 30%;
A second step (S122) of performing a convolution operation on the reduced image with a median filter having a size of 3x3; And
And rescaling the image to a 50% ratio of width and width again (S123).
The method of claim 3,
The second step (S122),
ID identification method having a low quality image, characterized in that the step of performing to improve the processing speed.
The method of claim 3,
The second step (S122),
And binarizing the identification card image.
3. The method of claim 2,
The pixel grouping step (S130),
A first step of setting the initial value of K (S131);
A second step (S132) of performing the K-means clustering algorithm;
A third step (S133) of obtaining the midpoints of all the pixels of the grouped clusters and comparing the distances between the midpoints and the identification image size to confirm the validity of K; And
If the K value is not valid, reducing the K by 1 and repeating the second step (S132) to the third step (S133); identification card having a low quality image comprising a .
3. The method of claim 2,
The area setting step (S140),
A storage step of storing four pixel positions corresponding to the minimum and maximum values of the X-axis, the minimum and maximum values of the Y-axis, and the coordinate plane which is the area of the grouped pixels (S141);
After the storing step, determining a validity of vertices of the remaining pixels by measuring a distance between the pixel of the minimum value of the X axis and the remaining pixels except the X axis (S142); And
And resetting the vertices by using the minimum and maximum values of the storing step when the vertices of the remaining pixels are not valid.
The method of claim 7, wherein
The reset step (S143),
And setting the minimum values of the X-axis and the Y-axis as vertices at the upper left, and the maximum values of the X-axis and the Y-axis as vertices at the lower right.
9. The method of claim 8,
The image extraction step (S150),
Extracting auxiliary images of four regions corresponding to minimum and maximum values of the X and Y axes (S151);
Setting a center point of the auxiliary image (S152);
A third step (S153) of calculating an angle for sub-image alignment of the four areas;
A fourth step (S154) of rotating the auxiliary image by the arrangement angle calculated in the third step;
A fifth step (S155) of calculating a black pixel of the X-axis and the Y-axis histogram after rotating the rotated image from -10 ° to 10 ° in the fourth step (S154);
And a sixth step (S156) of rotating the image to obtain an aligned image by applying the rotation angle having the largest number of rotations calculated in the fifth step (S155).
10. The method of claim 9,
The determination step (S160),
Performing template matching within a range of a predetermined area from an upper left corner of the aligned image,
The performing range of the template matching is,
The identification card identification method having a low quality image, characterized in that from the upper left of the aligned image to a position including the template image of the template matching target identification card.
The method of claim 10,
The template matching,
An ID card discrimination method having a low quality image, which uses NCC (Normalized Cross Correlation) and is performed in units of two pixels to shorten execution time.
The method of claim 10,
The image extraction step (S150),
And rotating the aligned image by 180 degrees to perform the template matching again.
12. The method of claim 11,
The template matching,
When the maximum value of the NCC is less than 0.7, the identification card identification method having a low quality image, characterized in that for performing template matching with a template of another identification image.

Images