Biometric identification methods, identification modules and chips

Abstract

This application discloses a biometric identification method, identification module, and chip. The biometric identification method includes: generating a sample feature model based on multiple sample biometric images acquired through multiple acquisitions; and determining whether the biometric image to be verified matches the sample feature model. If a match is found, the biometric image to be verified is determined to correspond to an identifiable user. The step of generating the sample feature model includes: determining matching image pairs based on feature points of the multiple sample biometric images, each matching image pair including a first image and a second image from the multiple sample biometric images, wherein the number of matching feature points between the first image and the second image is greater than a first preset value; determining the relative position parameters of each matching image pair in three-dimensional space; and establishing a three-dimensional sample feature model based on the relative position parameters of each matching image pair. This method eliminates duplicate feature points during the establishment of the sample feature model, improving the verification rate.

Description

Technical Field

[0001] This invention relates to the field of biometric identification technology, and in particular to a biometric identification method, identification module and chip. Background Technology

[0002] In recent years, biometric identification technology has been widely used in the consumer electronics field, with biometric identification being widely applied for identity verification. Examples include mobile phones, door locks, computers, and smartwatches. The key to biometric identification technology lies in the matching of biometric images.

[0003] Taking fingerprint recognition technology as an example, such as Figure 1 As shown, during fingerprint recognition, multiple fingerprint images 20 are first recorded and saved (i.e., registered fingerprint images 20). During recognition, the fingerprint image 10 to be verified is acquired and compared with each of the registered fingerprint images 20 to find the best match. However, with the development of thinner and smaller mobile smart terminals, the area of ​​the recorded fingerprint images is getting smaller and the number is increasing. Correspondingly, during matching, it often takes a long time to go through the registered fingerprint images 20, resulting in a slow recognition process. Summary of the Invention

[0004] In view of the above problems, the purpose of this application is to provide a biometric identification method, identification module, and chip. By establishing a three-dimensional model of the biometric feature based on the feature point information of the recorded biometric image, the fingerprint recognition speed and accuracy are improved.

[0005] According to one aspect of this application, a biometric identification method is provided, comprising: generating a sample feature model based on multiple sample biometric images acquired through multiple acquisitions; and determining whether a biometric image to be verified matches the sample feature model, wherein if they match, the biometric image to be verified corresponds to an identifiable user, wherein the step of generating the sample feature model based on multiple sample biometric images acquired through multiple acquisitions includes: determining matching image pairs based on feature points of the multiple sample biometric images, each matching image pair including a first image and a second image in the multiple sample biometric images, and the number of matching feature points between the first image and the second image being greater than a first preset value; determining the relative position parameters of each matching image pair in three-dimensional space; and establishing the three-dimensional sample feature model based on the relative position parameters of each matching image pair.

[0006] Optionally, the relative position parameter includes the rotation matrix and translation vector of the second image relative to the first image.

[0007] Optionally, the predicted position of the matching feature point in the second image is calculated based on the actual position of the matching feature point in the first image, the rotation matrix, and the translation vector. The rotation matrix and the translation vector are obtained by using the minimum value of the set of L2 norms of the differences between the actual position and the predicted position of each matching feature point in the second image as a constraint condition.

[0008] Optionally, establishing the three-dimensional sample feature model based on the relative position parameters of each of the matched image pairs includes: calculating a first distance parameter and a second distance parameter corresponding to each of the matched feature points based on the relative position parameters; calculating a first spatial position of each feature point based on the first distance parameter and the actual position of each of the matched feature points on the first image; calculating a second spatial position of each of the matched feature points based on the second distance parameter and the predicted position of each of the matched feature points on the second image; and calculating the spatial position of each of the matched feature points in the sample feature model based on the first spatial position and the second spatial position, wherein the first distance parameter is greater than zero and is used to represent the distance between the spatial position of any matched feature point and the corresponding position of the matched feature point on the first image, and the second distance parameter is greater than zero and is used to represent the distance between the spatial position of the matched feature point and the corresponding position of the matched feature point on the second image.

[0009] Optionally, the first distance parameter and the second distance parameter are obtained by using the minimum value of the set of 2-norms of the difference between the second spatial position and the first spatial position of each of the matching feature points as a constraint condition.

[0010] Optionally, the mean of the first spatial location and the second spatial location is calculated as the spatial location of the matching feature point in the sample feature model, and the first / second spatial location is calculated according to the following formula:

[0011] P1 = d j ×K -1 ×P j

[0012] P2=d i ×R ij ×K -1 ×P i +t ij

[0013] Wherein, P1 represents the first spatial position, P2 represents the second spatial position, and P i Let P be a vector in a 3×1 affine space. i The vector coordinates of P represent the positions of the feature points on the first image. j Let P be a vector in a 3×1 affine space.j The vector coordinates represent the position of the feature point on the second image, and in this affine space, P i vector coordinates and P j One of the vector coordinates is the same; R ij Let t be the third-order rotation matrix of the second image relative to the first image; ij Let d be the 3×1 translation vector of the second image relative to the first image. i >0 indicates the first distance parameter; d j >0 indicates the second distance parameter; K -1 It is the inverse matrix of the third-order image acquisition parameter matrix.

[0014] Optionally, the following formula represents the image acquisition parameter matrix:

[0015]

[0016] Where w represents the width of the sample image and h represents the height of the sample image.

[0017] Optionally, the method for determining matching image pairs includes image patch pairing.

[0018] Optionally, the method further includes repeating the step of generating a sample feature model based on multiple sample biometric images obtained from multiple acquisitions, until the sample feature model includes feature points with at least a second preset value, wherein the second preset value is greater than or equal to the first preset value.

[0019] According to another aspect of this application, a biometric identification module is provided, comprising: a processor for executing the calculation method as described in any of the preceding claims; and a memory for storing the three-dimensional biometric model and / or the feature point set of the three-dimensional biometric model.

[0020] According to a third aspect of this application, a biometric identification chip is provided, comprising: a biometric identification module as described above.

[0021] According to the biometric recognition method, recognition module, and chip provided in this application, duplicate feature points are removed during the construction of the three-dimensional biometric model, thereby reducing the traversal time of the image to be verified and accelerating the biometric matching recognition process. Furthermore, in this application, by restoring the spatial position of the biometrics, the constructed biometric model is more closely matched to the actual shape of the biometrics, further improving the accuracy of matching. Attached Figure Description

[0022] The above and other objects, features and advantages of the present invention will become more apparent from the following description of embodiments of the invention with reference to the accompanying drawings, in which:

[0023] Figure 1 A schematic diagram of ergonomic matching of biometric images is shown.

[0024] Figure 2 This diagram illustrates a biometric image stitching matching method.

[0025] Figure 3 A schematic flowchart illustrating the biometric identification method according to an embodiment of this application is shown;

[0026] Figure 4 This document illustrates a schematic flowchart of an embodiment of the present application for generating a sample feature model;

[0027] Figure 5A A schematic diagram of feature points in a single biometric image acquisition is shown.

[0028] Figure 5B A schematic diagram of feature points from two biometric images is shown.

[0029] Figure 5C A schematic diagram of feature points from three biometric images is shown.

[0030] Figure 6 This diagram illustrates the spatial transformation between the first and second images.

[0031] Figure 7 A schematic diagram showing the reconstruction of the spatial location of feature points based on the first and second images is shown.

[0032] Figure 8 This diagram illustrates a comparison of feature points for biometric identification in an embodiment of this application.

[0033] Figure 9 This diagram illustrates the structure of a biometric identification module according to an embodiment of this application. Detailed Implementation

[0034] Various embodiments of the invention will now be described in more detail with reference to the accompanying drawings. In the various drawings, the same elements are indicated by the same or similar reference numerals. For clarity, the various parts in the drawings are not drawn to scale.

[0035] It should be understood that certain terms have been used in the specification and prior claims to refer to specific components. Those skilled in the art will understand that manufacturers may use different names to refer to the same component. This specification and prior claims do not distinguish components based on differences in name, but rather on differences in function.

[0036] In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. Moreover, the word "connection" herein includes any direct and indirect electrical connection means. Therefore, if the text describes a first device connected to a second device, it means that the first device can be directly connected to the second device or indirectly connected to the second device through other means or connections.

[0037] It should be noted that the biometric information extraction method, feature recognition method, recognition module, and chip disclosed in this application are applicable to some biometric features with obvious characteristics, such as fingerprints, palm prints, toe prints, and sole prints. In the following description, fingerprint images are used as an example to illustrate the embodiments of this application, but this does not imply that the application is limited thereto.

[0038] Taking fingerprint recognition technology as an example, such as Figure 2 As shown, to improve the fingerprint recognition speed, in some embodiments, the recorded fingerprint image 20 is stitched together based on the overlapping portion 21 in the recorded fingerprint image 20 to obtain a stitched fingerprint image 30. During verification, the fingerprint image 10 to be verified is matched with the stitched fingerprint image 30, which can improve the fingerprint verification speed to a certain extent. However, since the finger itself is three-dimensional, and the fingerprint surface is approximately ellipsoidal, stitching two-dimensional features will inevitably cause deformation and loss of some feature information, thus affecting the accuracy of fingerprint recognition.

[0039] Figure 3 A schematic flowchart illustrating the biometric identification method according to an embodiment of this application is shown. Figure 3 As shown, the biometric identification method in this application specifically includes the following steps:

[0040] Step S10: Generate a sample feature model; and

[0041] Step S20: Determine whether the biometric image to be verified matches the sample feature model.

[0042] Specifically, Figure 4This diagram illustrates a schematic flowchart of the process for generating a sample feature model according to an embodiment of this application. Figure 4 As shown, the biometric identification method in this application specifically includes the following steps:

[0043] Step S11: Input multiple sample images of biometric features;

[0044] In this step, sample images of multiple biometric features of the fingerprint are recorded using sensors. These sensors include optical sensors, capacitive sensors, and others.

[0045] Step S12: Obtain the feature point set of each sample image;

[0046] In this step, feature points (e.g., bifurcation points, endpoints, island points, etc.) are extracted from each sample image to obtain the feature point set corresponding to each sample image. For example, if sample image An includes m feature points, then the feature point set P of sample image An is... n ={p n1 p n2 , ..., p nm}, where n and m are integers greater than or equal to 1.

[0047] Step S13: Determine matching image pairs based on the number of matching feature points that is the first preset value;

[0048] In this step, a sample image is selected as the first image, and a second image is obtained that is paired with the first image based on the matching of feature points. The number of matching feature points between the second image and the first image is at least a first preset value (e.g., 8).

[0049] Figure 5A A schematic diagram of feature points in a single biometric image acquisition is shown. Figure 5B A schematic diagram of feature points from two biometric images is shown. Figure 5C This diagram illustrates the feature points of three biometric images collected. (Example:) Figure 5A , 5B as well as Figure 5C As shown, by acquiring sample images of multiple fingerprint images, fingerprint feature points can be distinguished into feature points that were not acquired, feature points acquired once, and feature points acquired twice or more. Furthermore, as the number of acquisitions increases, more and more feature points are acquired twice or more. Therefore, depending on the number of sample images, some first images may not have a corresponding second image; some first images may have only one corresponding second image; and some first images may have multiple corresponding second images. Thus, the first preset value for matching feature points can be adjusted based on the number of acquired sample images to improve computational efficiency.

[0050] In this embodiment, to further shorten the time required for this step, a sample image can be selected as the first image, and then a second image matching the first image can be found among images that are similar to the first image using the image patch matching method.

[0051] Step S14: Obtain the relative position parameters of the matching image pair;

[0052] The first and second images each include at least a first preset number of matching feature points (e.g., 8). In this step, the relative position parameters of the first and second images are obtained using the first preset number of matching feature points. Using the first image as a reference pose, the relative position parameters include the rotation (rotation matrix) and translation (translation vector) of the second image relative to the first image.

[0053] Figure 6 A schematic diagram of the spatial transformation between the first and second images is shown, as follows: Figure 6 As shown in the embodiments of this application, an affine space is introduced to describe the position of feature points for ease of calculation. For any matching feature point in the first image and the second image, one bit of the vector coordinates in the affine space is the same. For example, in the affine space, for any matching feature point in the first image and the second image, the actual position of the feature point in the first image can be represented by the vector coordinates (x1, y1, 1) corresponding to a 3×1 vector, and correspondingly, the actual position of the feature point in the second image can be represented by the vector coordinates (x2, y2, 1) corresponding to a 3×1 vector.

[0054] Furthermore, taking the first image A i For the reference pose, the second image A j Relative to the first image A i The rotation matrix is ​​a third-order matrix R. ij The translation vector t is 3×1. ij .

[0055] So for the first image A i Second image A j Any matching feature point in the first image A i The position above is represented by a vector P in a 3×1 simulation space. i The vector coordinate representation in the second image A j The position above is represented by a 3×1 simulation space vector P. j The vector coordinate representation of . Then theoretically, there exists:

[0056] P j =R ij ×P i +t ij (1)

[0057] However, feature points are extracted from two-dimensional sample images based on the acquired fingerprint images. For the first image A... i Second image A j The matching feature points may not fully satisfy equation (1) above. Therefore, the rotation and translation are obtained according to the following optimization constraints:

[0058] min‖P j -(R ij ×P i +t ij )‖ (2)

[0059] Where ||x|| represents the L2 norm of x, and P i Let P be a vector in a 3×1 affine space. i The vector coordinates represent the positions of feature points on the first image; P j Let P be a vector in a 3×1 affine space. j The vector coordinates represent the positions of feature points on the second image; R ij t is the third-order rotation matrix of the second image relative to the first image; ij This is a 3×1 translation vector of the second image relative to the first image.

[0060] In other words, in this step, based on the matching feature point P i The actual position on the first image and the rotation matrix R ij Translation vector t ij Calculate the predicted position of the matching feature point in the second image to match the actual position P of the feature point in the second image. j The minimum value in the set of L2 norms of the difference between the predicted position and the actual position is used as a constraint to obtain the rotation and translation amounts.

[0061] It should be noted that, as described in step S13, in some embodiments, there may be a first image with multiple matching second images. Similarly, there may be several matching feature points in both the first image and the multiple second images. Therefore, when selecting feature points for calculating the transformation parameters between the first image and different second images, it is preferable to use feature points that are different in each second image. This improves the feature points of the 3D model while eliminating duplicate feature points, thereby increasing computational efficiency. If the second images do not have different feature points, then when calculating the transformation parameters, the optimal solution for rotation and translation is obtained based on the optimization constraints.

[0062] Step S15: Calculate the first distance parameter and the second distance parameter corresponding to the matching feature point based on the relative position parameter;

[0063] Figure 7This diagram illustrates the reconstruction of the spatial locations of feature points based on the first and second images. Figure 7 As shown, the first image A i The position above is P i Feature points and the second image A j The upper position is P j Let P be the corresponding feature point, and let P be the spatial location of that feature point. i The distance to P is the first distance parameter; P j The distance to P is the second distance parameter, based on P. i or P j P can be calculated for both, and theoretically:

[0064] P = d j ×K -1 ×P j =d i ×R ij ×K -1 ×P i +t ij (3)

[0065] Among them, P i Indicates the position of the feature point on the first image; P j Indicates the location of the feature point on the second image; P i Let P be a vector in a 3×1 affine space. i The vector coordinates of P represent the positions of the feature points on the first image. j Let P be a vector in a 3×1 affine space. j The vector coordinates represent the position of the feature point on the second image, and in the simulation space, P i vector coordinates and P j One of the vector coordinates is the same; R ij Let t be the third-order rotation matrix of the second image relative to the first image; ij Let d be the 3×1 translation vector of the second image relative to the first image. i >0 indicates the first distance parameter; d j >0 indicates the second distance parameter; K -1 This is the inverse matrix of the third-order image acquisition parameter matrix.

[0066]

[0067] Where w is the width of the sample image and h is the height of the sample image.

[0068] However, for the same reasoning as in step S14, feature points are extracted from two-dimensional sample images of the acquired fingerprint images. For the first image A...i Second image A j The matching feature points (e.g., P) i and P j It may not be possible to fully satisfy equation (3) above. Therefore, the P to P values ​​obtained first are based on the following optimization constraints. i and P j Distance:

[0069] min‖(d j ×K -1 ×P j )-(d i ×R ij ×K -1 ×P i +t ij )‖(5)

[0070] Where ||x|| represents the L2 norm of x, and P i Indicates the position of the feature point on the first image; P j Indicates the location of the feature point on the second image; P i Let P be a vector in a 3×1 affine space. i The vector coordinates of P represent the positions of the feature points on the first image. j Let P be a vector in a 3×1 affine space. j The vector coordinates represent the position of the feature point on the second image, and in the simulation space, P i vector coordinates and P j One of the vector coordinates is the same; R ij Let t be the third-order rotation matrix of the second image relative to the first image; ij Let d be the 3×1 translation vector of the second image relative to the first image. i >0 indicates the first distance parameter; d j >0 indicates the second distance parameter; K -1 It is the inverse matrix of the third-order image acquisition parameter matrix.

[0071] In other words, in this step, the first distance parameter d is obtained by using the set of minimum values ​​of the L2 norm of the difference between the predicted second spatial location and the first spatial location of each matched feature point as a constraint. i Second distance parameter d j The first spatial location is determined by the actual position P of each matched feature point on the first image. i and the first distance parameter d i The second spatial position is calculated based on the predicted position of each matched feature point on the second image and the second distance parameter d. j Obtained through calculation.

[0072] Step S16: Calculate the first spatial position or the second spatial position of each feature point based on the first distance parameter or the second distance parameter.

[0073] In this step, the first / second spatial position of each feature point is calculated according to the following formula (6) or (7) and the corresponding parameters.

[0074] P1 = d j ×K -1 ×P j (6)

[0075] P2=d i ×R ij ×K -1 ×P i +t ij (7)

[0076] Where P1 represents the first spatial position, P2 represents the second spatial position, and ideally P1 = P2. i Let P be a vector in a 3×1 affine space. i The vector coordinates of P represent the positions of the feature points on the first image. j Let P be a vector in a 3×1 affine space. j The vector coordinates represent the position of the feature point on the second image, and in the simulation space, P i vector coordinates and P j One of the vector coordinates is the same; R ij Let t be the third-order rotation matrix of the second image relative to the first image; ij Let d be the 3×1 translation vector of the second image relative to the first image. i >0 indicates the first distance parameter; d j >0 indicates the second distance parameter; K -1 It is the inverse matrix of the third-order image acquisition parameter matrix.

[0077] Step S17: Calculate the spatial position of each matching feature point in the sample feature model based on the first spatial position and the second spatial position:

[0078] In this step, for example, the average of the first and second spatial positions (i.e., the midpoint between the first and second spatial positions) is used as the spatial position of each matched feature point in the sample feature model. However, it should be understood that this application is not limited to this.

[0079] Step S18: Whether to acquire the spatial locations of feature points with a quantity equal to the second preset value;

[0080] In this step, it is determined whether the sample feature model includes feature points at least the second preset value. If the number of feature points in the sample feature model does not reach the second preset value, steps S13 to S17 are repeated, that is, the first image and the second image are replaced to obtain new matching feature points, and the spatial position of the new matching feature points is calculated, until the number of feature points in the sample feature model is greater than or equal to the second preset value. The second preset value should be greater than or equal to the first preset value.

[0081] In this application, feature points that have been collected two or more times are used to reconstruct the spatial location of biometric features, making the established biometric model more compatible with the actual shape of the biometric features and improving the accuracy of matching. Furthermore, duplicate feature points are deleted during the reconstruction process, which makes it easier to shorten the subsequent verification time and improve the verification efficiency.

[0082] Furthermore, Figure 8 This illustration shows a comparison diagram of feature points for biometric identification in an embodiment of this application. Figure 8 As shown, in step S20, the feature point set of the biometric sample feature model and the feature point set of the image to be verified are compared. If the number of matching feature points exceeds the preset number, the verification is successful; otherwise, the verification fails.

[0083] The present application provides a method for establishing and recognizing three-dimensional biometric models. During the establishment of the three-dimensional biometric model, duplicate feature points are removed, thereby reducing the traversal time of the image to be verified and accelerating the biometric matching and recognition process. Furthermore, in this application, by restoring the spatial position of the biometrics, the established biometric model is more closely matched to the actual shape of the biometrics, further improving the accuracy of the matching.

[0084] Furthermore, Figure 9 A schematic diagram of the biometric identification module provided in this application is shown. Figure 9 As shown, the biometric identification module 100 includes a processor 110 and a memory 120 communicatively connected thereto. The processor executes the biometric identification method provided in this application, and the memory stores the three-dimensional model created by the processor. Furthermore, this application also provides a biometric identification chip, including the biometric identification module 100. Since this biometric identification module and chip execute the aforementioned biometric identification method, they also have the beneficial effects described above, which will not be elaborated further here.

[0085] As described above, these embodiments of the present invention do not exhaustively describe all details, nor do they limit the invention to specific embodiments. Clearly, many modifications and variations can be made based on the above description. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to effectively utilize the invention and its modifications. The scope of protection of this invention should be determined by the scope defined in the claims of this invention.

Claims

1. A biometric identification method, comprising: Generate a sample feature model based on multiple biometric images of samples obtained through multiple collections; as well as The system determines whether the biometric image to be verified matches the sample feature model. If they match, the biometric image to be verified is determined to correspond to an identifiable user. The steps for generating a sample feature model based on multiple biometric images obtained from multiple acquisitions include: Matching image pairs are determined based on feature points of the plurality of sample biometric images. Each matching image pair includes a first image and a second image from the plurality of sample biometric images, and the number of matching feature points between the first image and the second image is greater than a first preset value. Determine the relative position parameters of each of the matched image pairs in three-dimensional space; and A three-dimensional sample feature model is established based on the relative position parameters of each matched image pair. The steps for determining the matching image pair include: Select one of the plurality of sample biometric images as the first image; and At least one second image that matches the first image is obtained, and each second image and the first image constitute a pair of matching images. The steps for determining the relative position parameters include: When both the first image and the plurality of second images include a plurality of matching feature points. If the multiple second images have different feature points, the relative position parameter is determined based on the different feature points.

2. The biometric identification method according to claim 1, wherein, The relative position parameters include the rotation matrix and translation vector of the second image relative to the first image.

3. The biometric identification method according to claim 2, wherein, The step of determining the relative position parameters further includes: If the multiple second images do not have different feature points, the predicted position of the matching feature point in the second image is calculated based on the actual position of the matching feature point in the first image, the rotation matrix, and the translation vector. The rotation matrix and the translation vector are obtained by using the minimum value of the L2 norm set of the differences between the actual position and the predicted position of each matching feature point in the second image as a constraint condition.

4. The biometric identification method according to claim 3, wherein, The step of establishing a three-dimensional sample feature model based on the relative position parameters of each matched image pair includes: Calculate the first distance parameter and the second distance parameter corresponding to each of the matching feature points based on the relative position parameters; and The first spatial position of each feature point is calculated based on the first distance parameter and the actual position of each matching feature point on the first image. Calculate the second spatial position of each matching feature point based on the second distance parameter and the predicted position of each matching feature point on the second image. Calculate the spatial position of each matching feature point in the sample feature model based on the first spatial position and the second spatial position. Wherein, the first distance parameter represents the distance between the spatial location of any matching feature point and the corresponding location of the matching feature point on the first image, and the second distance parameter represents the distance between the spatial location of the matching feature point and the corresponding location of the matching feature point on the second image.

5. The biometric identification method according to claim 4, wherein, The first distance parameter and the second distance parameter are obtained by using the minimum value of the set of 2 norms of the difference between the second spatial position and the first spatial position of each of the matching feature points as a constraint condition.

6. The biometric identification method according to claim 5, wherein, The mean of the first spatial position and the second spatial position is calculated as the spatial position of the matched feature point in the sample feature model. The first / second spatial positions are calculated according to the following formula: Wherein, P1 represents the first spatial position, P2 represents the second spatial position, and P i Let P be a vector in a 3×1 affine space. i The vector coordinates of P represent the positions of the feature points on the first image. j Let P be a vector in a 3×1 affine space. j The vector coordinates represent the position of the feature point on the second image, and in this affine space, P i vector coordinates and P j One of the vector coordinates is the same; R ij Let t be the third-order rotation matrix of the second image relative to the first image; ij Let d be the 3×1 translation vector of the second image relative to the first image. i >0 indicates the first distance parameter; d j >0 indicates the second distance parameter; K -1 It is the inverse of the third-order image acquisition parameter matrix.

7. The biometric identification method according to claim 6, wherein, The following formula represents the image acquisition parameter matrix: Where w represents the width of the sample image and h represents the height of the sample image.

8. The biometric identification method according to claim 1, wherein, The method for determining matching image pairs includes image patch matching.

9. The biometric identification method according to claim 1, wherein, It also includes repeating the step of generating a sample feature model from multiple sample biometric images obtained from multiple acquisitions, until the sample feature model includes feature points with at least a second preset value, wherein the second preset value is greater than or equal to the first preset value.

10. A biometric identification module, comprising: A processor for performing the computation method as described in any one of claims 1 to 9; as well as A memory for storing the three-dimensional model of the biometrics and / or the feature point set of the three-dimensional model of the biometrics.

11. A biometric identification chip, comprising: The biometric identification module as described in claim 10.

Images