A recognition method for low-quality fingerprints and incomplete fingerprints

Through image segmentation, enhanced processing and phase correlation calculation, combined with Gabor filter and mountain climbing search method, the matching difficulties of low-quality and incomplete fingerprints in matching feature points and small-area templates is solved, and more efficient fingerprint matching effect is achieved .

CN114898412BInactive Publication Date: 2024-12-03KUNMING UNIV OF SCI & TECH
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202210504420.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-05-10
Publication Date
2024-12-03
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Low-quality, incomplete fingerprints cannot be effectively matched during feature point matching, small-area template matching and pure phase correlation matching, which is prone to mismatch and the matching effect is not ideal.

Method used

Through image segmentation and enhancement processing, two-dimensional Fourier transform and Hough transform are used for correction, and Gabor filter is constructed for filtering. Combined with phase correlation calculation and improved mountain climbing search method, the maximum phase correlation coefficient of the small-area template is found to judge Fingerprint source.

Benefits of technology

The matching rate of low-quality and incomplete fingerprints is improved, the mismatch phenomenon is reduced, and the matching effect is significantly improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114898412B_ABST
    Figure CN114898412B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for identifying low-quality and incomplete fingerprints, and belongs to the field of image processing technology. First, the acquired low-quality and incomplete fingerprints are preprocessed, and then the fingerprint image is corrected by using two-dimensional Fourier transform and Hough transform, and the phase correlation between the corrected low-quality and incomplete fingerprints and the fingerprints in the database is calculated, and a threshold is set. When the phase correlation coefficient is higher than the threshold, the fingerprint image in the database is saved in a temporary database, and the approximate position of the small area template in the temporary database is found by the area matching method and expanded by T pixels around, and the expanded fingerprint image is saved in the temporary database 1; finally, the phase correlation between the small area template and each fingerprint to be matched in the temporary database 1 is calculated by the hill climbing search method, and the optimal solution is found, so as to determine the source of the low-quality and incomplete fingerprints. The present invention can effectively improve the phase correlation function peak value and recognition rate of low-quality and incomplete fingerprints, thereby reducing the risk of misidentification.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a method in the technical field of image processing, in particular to a recognition method for low-quality and incomplete fingerprints. Background Art

[0002] Fingerprints are one of the most widely used biometric technologies at present. With the rapid development of information technology, the application scenarios of fingerprints are becoming more and more extensive. For example, fingerprints are widely used in criminal investigation, and this application has also accelerated the progress of solving cases.

[0003] For low-quality and incomplete fingerprints: when using traditional feature point matching, it is impossible to find enough feature point pairs to make a successful match; when using small area template matching, the matching effect often does not achieve the expected effect, and even mismatching may occur; when using pure phase correlation matching, although the matching effect can be achieved, the phase correlation peak is still not ideal. Summary of the invention

[0004] The purpose of the present invention is to provide a method for identifying low-quality and incomplete fingerprints, which specifically comprises the following steps:

[0005] (1) Perform image segmentation and enhancement processing on the acquired low-quality and incomplete fingerprints: Calculate the intensity field, gradient field, direction field, and frequency field of the acquired fingerprint image information respectively, determine the fingerprint image according to the calculated intensity field and gradient field, segment the determined fingerprint image from the background, and perform equalization processing on the fingerprint image; construct the corresponding Gabor filter according to the local ridge direction and ridge frequency, and filter the local fingerprint image. The passband of the filter coincides with the local ridge direction and frequency distribution, which can enhance the local ridge direction and frequency information while filtering out other noise signals, thereby achieving fingerprint image enhancement.

[0006] (2) Correct the fingerprint image obtained in step (1) using two-dimensional Fourier transform and Hough transform.

[0007] (3) Calculate the phase correlation between the low-quality, incomplete fingerprint corrected in step (2) and the fingerprint image in the database.

[0008] (4) Preliminary screening: In the calculation process of step (3), the threshold is set to t, that is, the minimum phase correlation coefficient is t. When the phase correlation coefficient is greater than t, the fingerprint image is automatically stored in the temporary database until the database is traversed.

[0009] (5) According to the specific situation of the low-quality or incomplete fingerprint, a small-area template is selected for the corrected low-quality or incomplete fingerprint.

[0010] (6) Use the area matching method to find the approximate position of the small area template in the temporary database, and expand T pixels around the position as the center. The expanded fingerprint is the fingerprint to be matched, and the fingerprint image to be matched is saved in the temporary database 1.

[0011] (7) Use the improved hill climbing search method to calculate the maximum phase correlation coefficient between the small area template and each fingerprint image to be matched in the temporary database 1, find the optimal solution, and thus determine the source of low-quality and incomplete fingerprints.

[0012] Preferably, the step (2) of the present invention corrects the fingerprint image obtained in step (1), specifically comprising the following steps:

[0013] Assume that the original fingerprint image is f(x,y), size M×N, and the coordinate range is: x=0,1,2…M-1, y=0,1,2…N-1. Then the Fourier transform of the original fingerprint image is:

[0014]

[0015] Among them, M and N are the width and height of the original fingerprint image, respectively, and u and v are the frequency variables corresponding to x and y, respectively.

[0016] Then perform Hough transform on (1), the Hough transform is as follows:

[0017] In the rectangular coordinate system O-xy, suppose there is such a straight line, its distance from the origin O is l, and the direction angle is θ. Now let the coordinates of point A be (x, y), draw a straight line perpendicular to the x-axis through point A and intersecting the x-axis at point B, and draw a straight line perpendicular to OA through point B and intersecting OA at point C, so we have:

[0018] OC=x cosθ,CA=y sinθ (2)

[0019] but:

[0020] l=OC+CA=x cosθ+y sinθ (3)

[0021] Let D(x1,y1) be another coordinate point on the straight line. Draw a straight line perpendicular to AB through point D and intersect AB at point E. Draw a straight line perpendicular to AD through point E and intersect AD at point F. Then we have:

[0022] EF=(yy 1 )sinθ=(x 1 -x)cosθ (4)

[0023] Right now:

[0024] l=x cosθ+y sinθ=x 1cosθ+y 1 sinθ (5)

[0025] That is to say, every point on the straight line satisfies equation (5). From (4), we can know that:

[0026]

[0027] So the tilt angle of the original fingerprint image is:

[0028]

[0029] Then, the original fingerprint image is rotated and corrected according to the calculated tilt angle θ to obtain the corrected fingerprint image.

[0030] Preferably, the calculation (2) of the present invention and the phase correlation coefficient of the fingerprint image stored in the database specifically includes the following steps:

[0031] Assume that the size of fingerprint images f(x,y) and g(x,y) is M×N, and the coordinate range is: x=0,1,2,Λ,M-1, y=0,1,2,Λ,N-1.

[0032] If F(u,v) and G(u,v) are the two-dimensional Fourier transforms of f(x,y) and g(x,y), respectively, then:

[0033]

[0034]

[0035] Where: M, N are the width and height of the original fingerprint image, u, v are the frequency variables corresponding to x, y, respectively, θ G ,θ F is the phase angle, A F (u,v),A G (u,v) is the amplitude and j is an imaginary number.

[0036] Then the phase correlation spectrum of f(x,y) and g(x,y) in the spatial domain is:

[0037]

[0038] θ(u,v)=θ F (u,v)-θ G (u,v) (11)

[0039] Among them, G(u,v) * is the complex conjugate of G(u,v), and θ(u,v) is the phase difference.

[0040] If f(x, y) and g(x, y) come from the same fingerprint image, then equation (10) indicates that the phase difference between the two fingerprint images f(x, y) and g(x, y) is equal to the phase of the cross-power spectrum.

[0041] Performing inverse Fourier transform on equation (10), the phase correlation function is:

[0042]

[0043] Right now:

[0044] r fg (x,y)=δ(x 0 ,yy 0 ) (13)

[0045] and

[0046] In formula (13), r fg (x,y) is the 0 ,y 0 ) impulse function, if two fingerprint images are similar, their phase correlation function will have an obvious peak coefficient, the peak coefficient is called the maximum phase correlation coefficient, if the two images are not similar, the phase correlation coefficient is not obvious, the criterion for judging whether the fingerprint image matches is to judge the height of the phase correlation coefficient; through the position of the phase correlation coefficient, we can also find (x 0 ,y 0 ) position, further increasing or decreasing the displacement to reduce the slight error caused by the peak value.

[0047] Preferably, the improved hill climbing search method of the present invention is used to calculate the maximum phase correlation coefficient between the small area template and each fingerprint image to be matched in the temporary database 1, and find the optimal solution, so as to determine the source of low-quality and incomplete fingerprints. Taking any fingerprint image in the temporary database 1 as an example, the specific steps are as follows:

[0048] ① Take the center point of the fingerprint image to be matched as the starting point, and extract the small area template horizontally and vertically with the half width and half height of the small area template respectively.

[0049] ② Take the template where the center point of the fingerprint image to be matched is located as the starting template.

[0050] ③ Calculate the phase correlation coefficient between the small area template and the starting template, and assume that the phase correlation coefficient of the starting template is the maximum at this time, that is, C max .

[0051] ④ Taking the starting template as the center and the step length as T 1 (T 1≤T) pixels around the starting template to make 8 templates of the same size adjacent to the starting template, denoted as l i (x,y), and i=1,…,8, calculate the small area template and l i The phase correlation peak of (x,y) is denoted by C i , compare C i The size of the solution and find the optimal solution, denoted as r i , if r i <C max , the step size is reduced by 1 pixel, and 8 adjacent sub-region templates of the same size as the starting template are made in the direction of the starting template, denoted as l j (x,y), j = 1, ..., 8, calculate the small area template and l j The phase correlation peak of (x,y) is denoted by C j , find C in the same way j The better solution is denoted by r j , if r j <C max , repeat the above steps until the optimal solution is found; if r j >C max , and r j The corresponding template is the starting template, and the small area template and r are calculated. j The phase correlation peaks of the 8 adjacent sub-region templates around the corresponding template are denoted as C k , find C in the same way k The better solution is denoted by r k , if r k >r j , repeat the above steps until the optimal solution is found.

[0052] Beneficial effects of the present invention:

[0053] The present invention provides a small area phase correlation matching method specifically for low-quality and incomplete fingerprints. The matching method can effectively solve the defects caused by feature points, small area templates, and pure phase correlation matching, thereby improving the matching rate. Based on the above theory, in order to test the feasibility of the present invention, the present invention tested 100 pairs of fingerprint images and the detection was accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 The present invention is based on the identification principle diagram of low-quality and incomplete fingerprints;

[0055] Figure 2 The correction principle diagram of the present invention;

[0056] Figure 3 The present invention aims at extracting small-area template flow chart of low-quality and incomplete fingerprints;

[0057] Figure 4 Statistical graph of experimental test of the present invention. DETAILED DESCRIPTION

[0058] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited to the contents described above.

[0059] Example 1

[0060] A method for identifying low-quality and incomplete fingerprints, comprising the following steps:

[0061] (1) Perform image segmentation and enhancement processing on the acquired low-quality and incomplete fingerprints: Calculate the intensity field, gradient field, direction field, and frequency field of the acquired fingerprint image information respectively, determine the fingerprint image according to the calculated intensity field and gradient field, segment the determined fingerprint image from the background, and perform equalization processing on the fingerprint image; construct the corresponding Gabor filter according to the local ridge direction and ridge frequency, and filter the local fingerprint image. The passband of the filter coincides with the local ridge direction and frequency distribution, which can enhance the local ridge direction and frequency information while filtering out other noise signals, thereby achieving fingerprint image enhancement.

[0062] (2) Correct the fingerprint image obtained in step (1) using two-dimensional Fourier transform and Hough transform: Assume that the original fingerprint image is f(x, y), with a size of 360×400 pixels and a coordinate range of: x=0,1,2…359, y=0,1,2…399; then the Fourier transform of the original fingerprint image is:

[0063]

[0064] Then perform Hough transform on (1), the Hough transform is as follows:

[0065] In the rectangular coordinate system O-xy, suppose there is such a straight line, its distance from the origin O is l, and the direction angle is θ. Now let the coordinates of point A be (x, y), draw a straight line perpendicular to the x-axis through point A and intersecting the x-axis at point B, and draw a straight line perpendicular to OA through point B and intersecting OA at point C, so we have:

[0066] OC=x cosθ,CA=y sinθ (2)

[0067] but:

[0068] l=OC+CA=x cosθ+y sinθ (3)

[0069] Let D(x1,y1) be another coordinate point on the straight line. Draw a straight line perpendicular to AB through point D and intersect AB at point E. Draw a straight line perpendicular to AD through point E and intersect AD at point F. Then we have:

[0070] EF=(yy 1 )sinθ=(x 1 -x)cosθ (4)

[0071] Right now:

[0072] l=x cosθ+y sinθ=x 1 cosθ+y 1 sinθ (5)

[0073] That is to say, every point on the straight line satisfies equation (5). From (4), we can know that:

[0074]

[0075] So the tilt angle of the original fingerprint image is:

[0076]

[0077] Then, the original fingerprint image is rotated and corrected according to the calculated tilt angle θ to obtain the corrected fingerprint image.

[0078] (3) Calculate the phase correlation between the low-quality, incomplete fingerprint corrected in step (2) and the fingerprint image in the database:

[0079] Assume that the size of fingerprint images f(x,y) and g(x,y) is 360×400, and the coordinate range is: x=0,1,2…359, y=0,1,2…399;

[0080] If F(u,v) and G(u,v) are the two-dimensional Fourier transforms of f(x,y) and g(x,y), respectively, then:

[0081]

[0082]

[0083] Among them, A F (u,v),A G (u,v) is the amplitude, Phase

[0084] Then the phase correlation spectrum of f(x,y) and g(x,y) in the spatial domain is:

[0085]

[0086] θ(u,v)=θ F (u,v)-θ G (u,v) (11)

[0087] Among them, G(u,v) * is the complex conjugate of G(u,v), and θ(u,v) is the phase difference.

[0088] If f(x, y) and g(x, y) come from the same fingerprint image, then equation (10) indicates that the phase difference between the two fingerprint images f(x, y) and g(x, y) is equal to the phase of the cross-power spectrum.

[0089] Performing inverse Fourier transform on equation (10), the phase correlation function is:

[0090]

[0091] Right now:

[0092] r fg (x,y)=δ(x 0 ,yy 0 ) (13)

[0093] and

[0094] In formula (13), r fg (x,y) is the 0 ,y 0 ) impulse function, if two fingerprint images are similar, their phase correlation function will have an obvious peak coefficient, the peak coefficient is called the maximum phase correlation coefficient, if the two images are not similar, the phase correlation coefficient is not obvious, the criterion for judging whether the fingerprint image matches is to judge the height of the phase correlation coefficient; through the position of the phase correlation coefficient, we can also find (x 0 ,y 0 ) position, further increasing or decreasing the displacement to reduce the slight error caused by the peak value.

[0095] (4) Preliminary screening: In the calculation process of step (3), the threshold is set to t = 0.1, that is, the minimum phase correlation coefficient is 0.1. When the phase correlation coefficient is greater than 0.1, the fingerprint image is automatically stored in the temporary database until the database is traversed.

[0096] (5) According to the specific situation of the low-quality and incomplete fingerprint, a small-area template is selected for the corrected low-quality and incomplete fingerprint. In the process of selecting the small-area template, an area with relatively good fingerprint texture is selected as the small-area template. The size of the small-area template is 30×30 pixels. In this embodiment, the small-area template of the low-quality and incomplete fingerprint is selected as follows: Figure 3 shown.

[0097] (6) Use the area matching method to find the approximate position of the small area template in the temporary database. Take this position as the center and expand it 20 pixels in all directions. The expanded fingerprint is the fingerprint to be matched. At this time, the size of the fingerprint image to be matched is 70×70 pixels. The fingerprint image to be matched is saved in the temporary database 1.

[0098] (7) Use the improved hill climbing search method to calculate the maximum phase correlation coefficient between the small area template and each fingerprint image to be matched in the temporary database 1, and find the optimal solution to determine the source of low-quality and incomplete fingerprints. In this embodiment, in order to verify the feasibility of the proposed solution, 100 pairs of fingerprint images were tested. The test statistical results are as follows: Figure 4 shown.

[0099] The phase correlation coefficient of a small area is calculated according to the derivation process described in step (3). Taking any fingerprint image in the temporary database 1 as an example, the specific steps are as follows:

[0100] ① Taking the center point of the fingerprint image to be matched as the starting point, extract a small area template horizontally and vertically with the half width and half height of the small area template described in step (5) respectively.

[0101] ② Take the template where the center point of the fingerprint image to be matched is located as the starting template.

[0102] ③ Calculate the phase correlation coefficient between the small area template described in step (5) and the starting template in step ②, and assume that the phase correlation coefficient of the starting template is the maximum value at this time, that is, C max .

[0103] ④ With the starting template as the center, make 8 templates of the same size adjacent to the starting template with a step length of 5 pixels, denoted as l i (x,y), and i=1,…,8, calculate the small area template and l i The phase correlation peak of (x,y) is denoted by C i , compare C i The size of the solution and find the optimal solution, denoted as r i , if r i <C max , the step size is reduced by 1 pixel, and 8 adjacent sub-region templates of the same size as the starting template are made in the direction of the starting template, denoted as l j (x,y), j = 1, ..., 8, calculate the small area template and l j The phase correlation peak of (x,y) is denoted by C j , find C in the same way j The better solution is denoted by r j , if r j <C max, repeat the above steps until the optimal solution is found; if r j >C max , and r j The corresponding template is the starting template, and the small area template and r are calculated. j The phase correlation peaks of the 8 adjacent sub-region templates around the corresponding template are denoted as C k , find C in the same way k The better solution is denoted by r k , if r k >r j , repeat the above steps until the optimal solution is found.

[0104] In order to test the feasibility of the present invention, the present invention tested 100 pairs of fingerprint images. During the test, the present invention counted the peak values ​​of the phase correlation function. The statistical results are as follows: Figure 4 shown.

[0105] In the experiment, the key to judging whether the fingerprints are matched successfully lies in the phase correlation coefficient. The present invention uses t=0.1 as the threshold to judge whether two fingerprints are the same fingerprint. If the phase correlation coefficient is higher than the set threshold 0.1, it means that the two fingerprint images are from the same finger, otherwise the match fails. Figure 4 It can be seen that compared with the pure phase correlation matching algorithm, the peak value of the algorithm proposed in the present invention has changed to a great extent, and the matching rate has also been significantly improved.

Claims

1. A recognition method for low-quality and incomplete fingerprints. It is characterized in that The specific steps include the following: (1) Perform image segmentation and enhancement processing on the acquired low-quality and incomplete fingerprints: Calculate the intensity field, gradient field, direction field, and frequency field of the acquired fingerprint image information respectively, determine the fingerprint image based on the calculated intensity field and gradient field, segment the determined fingerprint image from the background, and perform equalization processing on the fingerprint image; construct a corresponding Gabor filter based on the local grain direction and grain frequency, and filter the local fingerprint image. The passband of the filter coincides with the local grain direction and frequency distribution, and can filter out other noise signals while enhancing the local grain direction and frequency information, thereby achieving fingerprint image enhancement; (2) Correcting the fingerprint image obtained in step (1) using two-dimensional Fourier transform and Hough transform; (3) calculating the phase correlation between the low-quality, incomplete fingerprint corrected in step (2) and the fingerprint image in the database; (4) Preliminary screening: In the calculation process of step (3), the threshold is set to t, that is, the minimum phase correlation coefficient is t. When the phase correlation coefficient is greater than t, the fingerprint image is automatically stored in the temporary database until the database is traversed; (5) According to the specific situation of the low-quality or incomplete fingerprint, select a small-area template for the corrected low-quality or incomplete fingerprint; (6) Use the area matching method to find the approximate position of the small area template in the temporary database, and expand T pixels around the position as the center. The expanded fingerprint is the fingerprint to be matched, and the fingerprint image to be matched is saved in the temporary database 1; (7) Use the improved hill climbing search method to calculate the maximum phase correlation coefficient between the small area template and each fingerprint image to be matched in the temporary database 1, find the optimal solution, and thus determine the source of low-quality and incomplete fingerprints.

2. The method for identifying low-quality and incomplete fingerprints according to claim 1, Features: In step (2), the fingerprint image obtained in step (1) is corrected, which specifically includes the following steps: Assume that the original fingerprint image is f(x,y), the size is M×N, and the coordinate range is: x=0,1,2…M-1, y=0,1,2…N-1; then the Fourier transform of the original fingerprint image is: Among them, M and N are the width and height of the original fingerprint image, respectively, and u and v are the frequency variables corresponding to x and y, respectively; Then perform Hough transform on equation (1), the specific steps are as follows: In the rectangular coordinate system O-xy, suppose there is such a straight line, its distance from the origin O is l, and the direction angle is θ. Now let the coordinates of point A be (x, y), draw a straight line perpendicular to the x-axis through point A and intersecting the x-axis at point B, and draw a straight line perpendicular to OA through point B and intersecting OA at point C, so we have: OC=xcosθ,CA=ysinθ (2) but: l=OC+CA=xcosθ+ysinθ (3) Let D(x 1 ,y 1 ) is another coordinate point on the straight line. D is a straight line perpendicular to AB and intersecting AB at point E. D is a straight line perpendicular to AD and intersecting AD at point F through point E. Then: EF=(y-y 1 )sinθ=(x 1 -x)cosθ (4) Right now: l=xcosθ+ysinθ=x 1 cosθ+y 1 sinθ (5) That is to say, every point on the straight line satisfies equation (5). From (4), we can know that: So the tilt angle of the original fingerprint image is: Then, the original fingerprint image is rotated and corrected according to the calculated tilt angle θ to obtain the corrected fingerprint image.

3. The method for identifying low-quality and incomplete fingerprints according to claim 1, Features: The low-quality, incomplete fingerprint corrected in step (2) is subjected to phase correlation calculation with the fingerprint image in the database. The specific steps are as follows: Assume that the size of fingerprint images f(x,y) and g(x,y) is M×N, and the coordinate range is: x=0,1,2,Λ,M-1, y=0,1,2,Λ,N-1; If F(u,v) and G(u,v) are the two-dimensional Fourier transforms of f(x,y) and g(x,y), respectively, then: Among them, M and N are the width and height of the original fingerprint image, u and v are the frequency variables corresponding to x and y, respectively, and θ G ,θ F is the phase angle, A F (u,v),A G (u,v) is the amplitude; j is an imaginary number; Then the phase correlation spectrum of f(x,y) and g(x,y) in the spatial domain is: θ(u,v)=θ F (u,v)-θ G (u,v) (11) Where G(u,v)* is the complex conjugate of G(u,v), and θ(u,v) is the phase difference; If f(x, y) and g(x, y) are derived from the same fingerprint image, then equation (10) indicates that the phase difference between the two fingerprint images f(x, y) and g(x, y) is equal to the phase of the cross-power spectrum; Performing inverse Fourier transform on equation (10), the phase correlation function is: Right now: r fg (x,y)=δ(xx 0 ,yy 0 ) (13) and In formula (13), r fg (x,y) is the 0 ,y 0 ) impulse function, if two fingerprint images are similar, their phase correlation function will have an obvious peak coefficient, the peak coefficient is called the maximum phase correlation coefficient, if the two images are not similar, the phase correlation coefficient is not obvious, the criterion for judging whether the fingerprint image matches is to judge the height of the phase correlation coefficient; through the position of the phase correlation coefficient, we can also find (x 0 ,y 0 ) position, further increasing or decreasing the displacement to reduce the slight error caused by the peak value.

4. The method for identifying low-quality and incomplete fingerprints according to claim 1, Features: The specific steps of the hill climbing search method are: ① Take the center point of the fingerprint image to be matched as the starting point, and extract the small area template horizontally and vertically with the half width and half height of the small area template respectively; ② Take the template where the center point of the fingerprint image to be matched is located as the starting template; ③ Calculate the phase correlation coefficient between the small area template and the starting template, and assume that the phase correlation coefficient of the starting template is the maximum at this time, that is, C max ; ④ Taking the starting template as the center and the step length as T 1 (T 1 ≤T) pixels around the starting template to make 8 templates of the same size adjacent to the starting template, denoted as l i (x,y), and i=1,…,8, calculate the small area template and l i The phase correlation peak of (x,y) is denoted by C i , compare C i The size of the solution and find the optimal solution, denoted as r i , if r i <C max , the step size is reduced by 1 pixel, and 8 adjacent sub-region templates of the same size as the starting template are made in the direction of the starting template, denoted as l j (x,y), j = 1, ..., 8, calculate the small area template and l j The phase correlation peak of (x,y) is denoted by C j , find C in the same way j The better solution is denoted by r j , if r j <C max , repeat the above steps until the optimal solution is found; if r j >C max , and r j The corresponding template is the starting template, with T 1 Calculate the small area template and r for the step size j The phase correlation peaks of the 8 adjacent sub-region templates around the corresponding template are denoted as C k , find C in the same way k The better solution is denoted by r k , if r k >r j , repeat the above steps until the optimal solution is found.

Citation Information

Patent Citations

  • Method for calculating area of fingerprint overlapping region and electronic device

    CN105205439A

  • Banknote forming mode inspection and identification method based on banknote printing pattern characteristics

    CN111310628A

  • Image data comparison method and device

    CN111797921A

  • Incomplete region processing method for low-quality fingerprint image

    CN112115848A

  • Image positioning method and device based on NCC algorithm

    CN113643370A