TWI553501B - Iris feature identification method and its system - Google Patents

Description

Iris feature identification method and system thereof

The invention relates to an identification method and a device thereof, in particular to a method and system for utilizing iris feature analysis for identification.

Since ancient times, identity has been an important means of protecting people or materials, and extremely important personnel or materials can only be identified or read by approved individuals after being identified by layers. In the prior art, the specific account number and password are used for identification, or the biometrics of the user are read for identification. Biometric technology uses individual biobehavioral features such as DNA, facial features, eye iris, hand sweat, veins, voiceprints, fingerprints, signatures or typing speeds for identification.

When the iris is used for identification, the existing technical means uses an image capturing device to acquire iris image data, and performs feature comparison analysis on the iris image data. The existing iris image data feature comparison analysis, as described in the "Iris Recognition System" patent of the Republic of China Patent No. I335544, first locates the pupil center, and selects a first pixel group of the first circumference with a first radius and After the second pixel group of the second circumference is selected by a second radius, the iris region of the eye image is located, and the iris region is normalized and equalized to generate an equalized iris region. The features of the iris region are extracted to perform feature comparison with the iris image data in the database for identity authentication. In the case of the No. I335544 patent, whether the center of the pupil is located or the iris region is positioned, the two-pixel group of two circles is selected with different radii, and the difference between the total of the two pixel groups is calculated. Row positioning. However, this method will increase the complexity of image conversion, increase the burden on the processor, and is not conducive to reducing the overall recognition and comparison time.

In view of the increase in image conversion complexity caused by the existing iris recognition system, the main object of the present invention is to provide an iris feature recognition method and system with high efficiency and high recognition effect, thereby reducing the burden on the processor and reducing the overall recognition and The time of comparison.

In order to achieve the above object, the main technical means adopted by the present invention is that the iris feature recognition system includes: a storage device, an image data capture device, and a process electrically connected to the storage device and the image data capture device. Device. At least one iris image data sample is stored in the storage device. The image capturing device captures an iris image data. The processor receives the iris image data, first performs grayscale processing on the iris image data, determines a pupil region, and further selects at least one feature line segment tangent to the pupil region from the iris image data, and then according to The gray line of the feature line segment is selected from the feature line segment to select at least one feature point, and an identification area is respectively set according to the feature point, and the identification area is compared with a gray image similarity of the iris image data sample; When the gray-scale similarity comparison is the same, the identity authentication result is confirmed to be successful; when the gray-scale similarity comparison is different, the identity authentication result is confirmed to be unsuccessful.

The technical method of the invention further comprises the following steps: pre-setting at least one iris image data sample, extracting an iris image data, performing gray scale processing on the iris image data, and determining one In the pupil region, at least one feature line segment tangential to the pupil region is selected from the iris image data, and at least one feature point is selected according to the gray scale distribution of the feature line segment, and a feature point is respectively set according to the feature point Identifying the area, and comparing the identification area with the iris image data sample for gray scale similarity; when the gray level similarity comparison is the same, confirming that the identity authentication result is successful; when the gray level similarity comparison is different, confirming the identity The authentication result failed.

The present invention defines the iris area by extracting the iris image data according to the above method, and compares the identification area with the iris image data sample for gray scale similarity to confirm whether the captured iris image data is Correct to authenticate. The invention adopts different image matching feature region capturing methods to reduce the complexity of image conversion, reduce the burden on the processor and identify the time of comparison.

10‧‧‧Storage device

20‧‧‧Image capture device

30‧‧‧ Processor

40‧‧‧ display device

50‧‧‧ pupil area

60‧‧‧Characteristic line segments

61‧‧‧ feature points

62‧‧‧ Identification area

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a block diagram of a preferred embodiment of the iris feature recognition system of the present invention.

2 is a flow chart of a preferred embodiment of the iris feature identification method of the present invention.

3 is a schematic diagram of selecting three feature line segments in accordance with a preferred embodiment of the present invention.

4 is a schematic view showing a display screen of iris characteristic data according to a preferred embodiment of the present invention.

FIG. 5 is a schematic diagram of selecting six identification regions according to a preferred embodiment of the present invention.

Figure 6 is a schematic illustration of an identification area in accordance with a preferred embodiment of the present invention.

The technical means adopted by the present invention for achieving the intended purpose are further explained below in conjunction with the drawings and preferred embodiments of the present invention.

Referring to FIG. 1 and FIG. 2, the present invention is an iris feature recognition method and system thereof. The iris feature recognition system includes a storage device 10, an image capture device 20, and a processor 30. The storage device 10 stores an iris image data sample, and the iris data image sample is used as a comparison standard for identity authentication, that is, a recognition area of a pre-acquired iris image data of a user having a qualified authentication identity, and Store it. The image capturing device 20 is configured to capture an iris image data of the person to be identified. The processor 30 is electrically connected to the storage device 10 and the image capturing device 20 for receiving The iris image data captured by the image capturing device 20 is grayscaled, and a pupil region 50 is determined, and at least one of the iris image data is selected to be tangent to the pupil region 50. The feature line segment 60, for example, the pupil region 50 is a circle, and the line of the circle is selected as the feature line segment 60. In the preferred embodiment, the processor 30 performs binarization processing on the grayscale iris image data to determine the pupil region 50, and the processor 30 selects three feature segments, as shown in FIG. 3. As shown, the three feature line segments 60 are obtained by first obtaining a square, and one of the four sides of the square is tangent to the pupil region 50, and the upper side of the square is removed to eliminate the shadow of the upper eyelashes of the human eye. The error is the remaining line segment, which is the three feature line segment 60.

As shown in FIG. 4 , after obtaining the three feature line segments 60 , the processor 30 selects at least one feature point in each feature line segment 60 according to the grayscale processed iris image data. In the preferred embodiment, in the three feature line segments 60, two feature points 61 having the lowest gray level and the second lowest level are selected for each feature line segment 60, for a total of six feature points 61. The three-fold line graph in FIG. 4 is a gray scale degree distribution diagram of the three characteristic line segments 60, wherein the horizontal axis is the length coordinate of the three characteristic line segments, and the unit is a pixel, and the vertical axis thereof is a gray scale degree. Referring to FIG. 5, the processor 30 selects an identification area 62 for each of the feature points 61 on the iris image data, and a total of six identification areas 62. In the preferred embodiment, as shown in FIG. 6, the image length of the three feature line segments 60 is 90 pixels. The identification area 62 is a square having a length of 30 pixels, and the identification area 62 is selected by the feature point 61 to be offset from the pupil area 50 by 5 pixels, and then to the characteristic line segment 60 having the feature point 61, respectively. The parallel two opposite directions are offset by 15 pixels each to define the first side of the identification area 62, and are further reversely shifted by 25 pixels from the feature point 61, and then respectively to the characteristic line segment 60 having the feature point 61. The parallel two opposite directions are offset by 15 pixels each to define a second side of the identification area 62, and the opposite first side is connected with two corresponding end points of the second side to form the side length of 30 pixels. Square identification area 62.

The processor 30 further compares the six identification regions 62 with the iris image data samples by grayscale similarity; when the grayscale similarity comparison is the same, the identity authentication result is confirmed to be successful; when the grayscale similarity comparison is different When confirming the identity verification result failed. Further, when the processor 30 performs gray scale similarity comparison on the identification regions 62 of the iris image data samples, when the gray scale similarity ratios of the n identification regions are the same, Setting a comparison score to m, and n and m are positive correlations, for example, n=m, that is, when the grayscale similarity ratio of one recognition region 62 is the same, the comparison score is set to 1, and When the grayscale similarity of the two recognition regions 62 is the same, the comparison score is set to 2, and so on, the alignment score is set to at most 6, and the processor 30 presets a standard score, and When the gray scale similarity of the six identification regions 62 is compared, the comparison score is compared with the standard score. If the comparison score is greater than or equal to the standard score, the identity authentication result is determined to be successful, if the ratio is successful. When the score is less than the standard score, the authentication result is determined to have failed. For example, the preset standard score is 4, and the gray-scale similarity comparison result of the five recognition regions 62 is the same, so the comparison score is 5, and the current identity is The result of the certification is a success. In the preferred embodiment, the gray scale similarity comparison uses a correlation coefficient method.

Further, the iris feature recognition system further includes a display device 40 electrically connected to the processor 30, and the processor 30 transmits the iris image data to the display device 40 for display.

The iris image data sample is an identity authentication comparison standard that is collected and stored in advance by performing a sampling procedure before using the present invention for identity authentication. The sampling program captures iris image data of a user with qualified identity authentication by the image capturing device 20, and the processor 30 obtains the feature region of the iris image data according to the same manner as described above. The six feature areas of the iris image data of the qualified identity authenticated user are stored in the storage device as the iris image data sample.

Referring to FIG. 2, the iris feature identification method of the present invention includes the following steps: preliminarily providing at least one iris image data sample (S10); Extracting at least one iris image data (S11); performing grayscale processing on the iris image data (S12); determining a pupil region (S13); selecting at least one characteristic line segment tangent to the pupil region from the iris image data 60 (S14); according to the grayscale distribution of the grayscaled feature line segment 60, at least one feature point 61 is selected in the feature line segment 60 (S15); an identification region 62 is set according to the feature point 61 (S16); The identification area 62 performs gray scale similarity comparison with the iris image data sample (S17); when the gray level similarity comparison is the same, the identity authentication result is confirmed to be successful (S18); when the gray level similarity comparison is different, Confirm that the authentication result failed (S19).

In the preferred embodiment, in the above step S14, the three feature line segments 60 are selected, and the selection of the three feature line segments 60 is performed by binarizing the grayscale iris image data to determine a pupil. The region 50, and a square in which one side is tangent to the pupil region 50, and after removing the upper side of the square, the remaining three sides are the three feature line segments 60. In the above step S15, two feature points 61 are respectively selected in each feature line segment 60, and the two feature points 61 of the same feature line segment 60 are in the gray-scaled feature line segment 60, and the gray level is selected to be the lowest. The second lowest point is the two feature points 61. The identification area 62 is a square having a length of 30 pixels, and the identification area 62 is selected by the feature point 61 to be offset from the pupil area 50 by 5 pixels, and then to the characteristic line segment 60 having the feature point 61, respectively. The parallel two opposite directions are offset by 15 pixels each to define the first side of the identification area 62, and are further reversely shifted by the feature point 61 to the pupil area 50 by 25 pixels, and then respectively have the feature point The characteristic line segments 60 of the 61 are offset in opposite directions by 15 pixels each to define a second side of the identification region 62, and the opposite sides of the opposite first and second sides are connected to form the side A square recognition area 62 of 30 pixels long.

When the step of performing the gray-scale similarity comparison on the identification region of the iris image data sample is performed (S17), the method further includes: presetting a standard score (S171); and graying the n identification regions. When the order similarity is the same, a pairwise score is set to m, where n is positively correlated with m, and when the grayscale similarity of each identified region is compared, the comparison score is compared with the standard score ( S172); if the comparison score is greater than or equal to the standard score, confirming that the identity authentication result is successful (S18); if the comparison score is less than the standard score, confirming that the identity authentication result fails (S19).

In general, in the iris of the human eye, the iris characteristics near the lowest point of the gray scale are most obvious. Therefore, the present invention first selects the three feature line segments 60, and then performs grayscale processing on the iris image to find the lowest gray point degree and the second low point two feature points 61 in each feature line segment 60, and further based on the The identification area 62 near the feature point 61 performs gray scale similarity comparison for identity authentication. By extracting the iris features in different ways, and without complicated image conversion, the burden on the processor 30 is reduced, and the complexity of image conversion is reduced, which can further save the overall identification and comparison time.

The above is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Although the preferred embodiments of the present invention are disclosed above, the present invention is not limited thereto, and is generally Those skilled in the art can make some modifications or modifications to equivalent embodiments using the above-disclosed technical contents without departing from the technical scope of the present invention, but the present invention does not deviate from the technical solution of the present invention. Technical Substantials Any simple modifications, equivalent changes and modifications made to the above embodiments are still within the scope of the technical solutions of the present invention.

Claims (12)

An iris feature recognition system includes: a storage device storing at least one iris image data sample; an image capturing device for capturing an iris image data; a processor electrically connected to the storage device and the image capturing device The device receives the iris image data, and performs grayscale processing on the iris image data, determines a pupil region, and selects three feature segments that are tangent to the pupil region, and selects two features in each feature segment Pointing, and selecting an identification area according to the position of each feature point, and performing gray scale similarity comparison on the identification area and the iris image data sample; when the gray level similarity comparison is the same, confirming that the identity authentication result is successful; When the gray-scale similarity comparison is different, the identity authentication result is confirmed to be unsuccessful; wherein the processor selects a three-character line segment, and selects a square, and one of the four sides of the square is tangent to the pupil region, and is removed. After the upper side of the square, the remaining three sides are respectively the feature line segments; wherein the processor selects the same feature line segment When the feature points, tied to the characteristics of the iris image data segment after grayscale processing selected gray scale level of the lowest and the second lowest of two feature points. The iris feature recognition system of claim 1, wherein the processor further presets a standard score, and performs gray scale similarity comparison on each of the identification regions and the identification region of the iris image data sample, respectively. When the gray-scale similarity ratios of the identification regions are the same, a comparison score is set to m, where n is positively correlated with m, and the gray-scale similarity of each recognition region is compared, and the comparison score is obtained. Compared with the standard score, if the comparison score is greater than or equal to the standard score, the identity authentication result is determined to be successful, and if the comparison score is less than the standard score, the identity authentication result is determined to be unsuccessful. The iris feature recognition system of claim 1, wherein the processor performs binarization processing on the grayscale iris image data to determine the pupil region. The iris feature recognition system according to claim 1, wherein the gray scale similarity comparison system adopts a correlation coefficient method. The iris feature recognition system according to any one of claims 1 to 4, wherein the feature line segment has a length of 90 pixels. The iris feature recognition system according to any one of claims 1 to 4, wherein the identification area is a square having a side length of 30 pixels, and the processor selects the identification area from the feature point to the pupil The region is offset by 5 pixels, and then the first side of the identification region is defined by offsetting 15 pixels from the opposite directions of the two parallel to the feature line segment having the feature point, and the processor re-directs the feature from the feature point to the pupil The region is reversely offset by 25 pixels, and each of the two pixels is offset from the opposite direction of the feature line segment having the feature point by 15 pixels, respectively, to define a second side of the identification region, and the opposite first side is The two corresponding ends of the second side are connected to form a square identification area having a side length of 30 pixels. An iris feature identification method includes: presetting at least one iris image data sample; extracting an iris image data; performing grayscale processing on the iris image data; determining a pupil region; selecting three from the iris image data a characteristic line segment tangent to the pupil region; selecting two feature points in each of the feature line segments according to the gray scale distribution of the grayscaled feature line segments; setting an identification region according to each of the feature points; the identification region and the iris The image data samples are compared with the gray scale similarity; when the gray scale similarity comparison is the same, the identity authentication result is confirmed to be successful; when the gray scale similarity comparison is different, the identity authentication result is confirmed to be invalid; In the step of selecting three characteristic line segments tangent to the pupil region from the iris image data, a square is selected, and one of the four sides of the square is tangent to the pupil region, and the upper side of the square is removed, The lower three sides are the three characteristic line segments; wherein in the step of selecting two feature points in each of the characteristic line segments according to the gray scale distribution of the grayscaled characteristic line segments, the iris is replaced by the gray scaled processing Two feature points with the lowest gray level and the second lowest are selected in the characteristic line segments of the image data. The method for identifying an iris feature according to claim 7, wherein the step of comparing the grayscale similarity between the identification region and the identification region of the iris image data sample further comprises: presetting a standard score; When the gray-scale similarity of the n identification regions is the same, set a comparison score to m, where n and m are positively correlated, and when the gray-scale similarity of each recognition region is compared, the comparison is compared. The score is equal to the standard score; if the comparison score is greater than or equal to the standard score, the identity verification result is confirmed to be successful; if the comparison score is less than the standard score, the confirmation identity verification result fails. The iris feature identification method according to claim 7, wherein the pupil region is determined according to binarization processing on the grayscale iris image data. The iris feature identification method according to claim 7, wherein the gray scale similarity comparison system adopts a correlation coefficient method. The iris feature recognition method according to any one of claims 7 to 10, wherein the length of each feature line segment is 90 pixels. The iris feature identification method according to any one of claims 7 to 10, wherein in the step of respectively setting an identification area according to each feature point, the identification area is a square having a side length of 30 pixels, and the identification area is Offset the pupil area by 5 pixels from the feature point, and then respectively The feature line segments of the feature point are offset by 15 pixels in opposite directions to define a first side of the identification region, and are further reversely shifted by 25 pixels from the feature point to the pupil region, and respectively The feature line segments of the feature points are offset by two pixels in each of the opposite directions to define a second side of the identification region, and the opposite sides of the first side and the second side are connected to form the side length. A square recognition area of 30 pixels.