US-12619697-B2 - Identity recognition method and apparatus, and feature extraction method and apparatus for biometric pattern information

US12619697B2US 12619697 B2US12619697 B2US 12619697B2US-12619697-B2

Abstract

An identity recognition method includes: acquiring biometric pattern information describing a biometric pattern of a first object; performing feature extraction on the biometric pattern information to obtain a global pattern feature and a local pattern feature; fusing the global pattern feature and the local pattern feature to obtain a fused pattern feature of the first object; and performing identity recognition on the first object based on the fused pattern feature of the first object.

Inventors

Lei Shen
Chuhan ZHOU
Ruixin ZHANG
Kai Zhao
Tao Wang
Yingyi ZHANG
Shouhong DING

Assignees

TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED

Dates

Publication Date: 20260505
Application Date: 20231103
Priority Date: 20220209

Claims (16)

1 . An identity recognition method, performed by a computer device and comprising: acquiring biometric pattern information describing a biometric pattern of a first object, the biometric pattern information including palm pattern image information; obtaining a plurality of finger gap key points based on the biometric pattern information, wherein the plurality of finger gap key points comprise a first key point, a second key point, and a third key point, the first key point is between an index finger and a middle finger of a palm, the second key point is between the middle finger and a ring finger of the palm, and the third key point is between the ring finger and a little finger of the palm; establishing a rectangular coordinate system in the biometric pattern information, wherein a direction from the third key point to the first key point is a positive direction of an x-axis of the rectangular coordinate system, a direction from an intersection point where a line perpendicular to the x-axis passing through the second key point intersects the x-axis to the second key point as a positive direction of a y-axis of the rectangular coordinate system; determining a central point on the y-axis in a negative y-axis direction, wherein the central point is located on the negative y-axis direction at a particular ratio of a distance from the first key point to the third key point; determining a region of interest in the biometric pattern information based on the central point, the central point being at a center of the region of interest; extracting information in the region of interest to obtain updated biometric pattern information; performing feature extraction on the updated biometric pattern information to obtain a global pattern feature and a local pattern feature, the global pattern feature describing a global feature of the updated biometric pattern information of the first object, and the local pattern feature describing a local feature of the updated biometric pattern information of the first object; fusing the global pattern feature and the local pattern feature of the first object to obtain a fused pattern feature of the first object; and performing identity recognition on the first object based on the fused pattern feature of the first object by comparing the fused pattern feature of the first object with a sample pattern feature to obtain an identity determination result of the first object, the sample pattern feature being used for describing a biometric pattern information feature of a target object, wherein comparing the fused pattern feature of the first object with the sample pattern feature comprises: calculating a cosine similarity between the fused pattern feature and the sample pattern feature; obtaining, in response to the cosine similarity satisfying an identity determination threshold, the identity determination result that the first object is the target object; and obtaining, in response to the cosine similarity not satisfying the identity determination threshold, the identity determination result that the first object is not the target object.
2 . The method according to claim 1 , wherein performing the feature extraction on the biometric pattern information to obtain the global pattern feature and the local pattern feature comprises: performing hidden layer feature extraction on the biometric pattern information to obtain a texture hidden layer feature of the first object, the texture hidden layer feature describing a hidden layer representation of target palmprint information; performing feature extraction on the texture hidden layer feature by a first feature extraction method to obtain the global pattern feature of the first object, the first feature extraction method including a feature dimensionality reduction-based feature extraction method; and performing feature extraction on the texture hidden layer feature by a second feature extraction method to obtain the local pattern feature of the first object, the second feature extraction method including a feature receptive field increase-based feature extraction method.
3 . The method according to claim 2 , wherein performing the feature extraction on the texture hidden layer feature by the second feature extraction method to obtain the local pattern feature of the first object comprises: respectively processing the texture hidden layer feature by n second feature extraction methods to obtain n local sub-features of the first object, n being a positive integer, and feature receptive fields of the n second feature extraction methods being different; and splicing the n local sub-features to obtain the local pattern feature of the first object.
4 . The method according to claim 1 , comprising: performing self-attention mechanism processing on the local pattern feature to obtain an updated local pattern feature, wherein fusing the global pattern feature and the local pattern feature of the first object to obtain the fused pattern feature of the first object comprises: fusing the global pattern feature and the updated local pattern feature of the first object to obtain the fused pattern feature of the first object.
5 . The method according to claim 4 , wherein the self-attention mechanism processing comprises an activation process and a regularization process; and performing the self-attention mechanism processing on the local pattern feature to obtain the updated local pattern feature comprises: activating the local pattern feature to obtain a second activated feature; regularizing the local pattern feature to obtain a second regularized feature; and multiplying the second activated feature by the second regularized feature to obtain the updated local pattern feature of the first object.
6 . The method according to claim 1 , further comprising: performing interpolation for the biometric pattern information.
7 . A computer device, comprising: one or more processors and a memory, the memory storing at least one program, and the one or more processor being configured to execute the at least one program to implement an identity recognition method by performing: acquiring biometric pattern information describing a biometric pattern of a first object, the biometric pattern information including palm pattern image information; obtaining a plurality of finger gap key points based on the biometric pattern information, wherein the plurality of finger gap key points comprise a first key point, a second key point, and a third key point, the first key point is between an index finger and a middle finger of a palm, the second key point is between the middle finger and a ring finger of the palm, and the third key point is between the ring finger and a little finger of the palm; establishing a rectangular coordinate system in the biometric pattern information, wherein a direction from the third key point to the first key point is a positive direction of an x-axis of the rectangular coordinate system, a direction from an intersection point where a line perpendicular to the x-axis passing through the second key point intersects the x-axis to the second key point as a positive direction of a y-axis of the rectangular coordinate system; determining a central point on the y-axis in a negative y-axis direction, wherein the central point is located on the negative y-axis direction at a particular ratio of a distance from the first key point to the third key point; determining a region of interest in the biometric pattern information based on the central point, the central point being at a center of the region of interest; extracting information in the region of interest to obtain updated biometric pattern information; performing feature extraction on the updated biometric pattern information to obtain a global pattern feature and a local pattern feature, the global pattern feature describing a global feature of the updated biometric pattern information of the first object, and the local pattern feature describing a local feature of the updated biometric pattern information of the first object; fusing the global pattern feature and the local pattern feature of the first object to obtain a fused pattern feature of the first object; and performing identity recognition on the first object based on the fused pattern feature of the first object by comparing the fused pattern feature of the first object with a sample pattern feature to obtain an identity determination result of the first object, the sample pattern feature being used for describing a biometric pattern information feature of a target object, wherein comparing the fused pattern feature of the first object with the sample pattern feature comprises: calculating a cosine similarity between the fused pattern feature and the sample pattern feature; obtaining, in response to the cosine similarity satisfying an identity determination threshold, the identity determination result that the first object is the target object; and obtaining, in response to the cosine similarity not satisfying the identity determination threshold, the identity determination result that the first object is not the target object.
8 . The computer device according to claim 7 , wherein the one or more processors are further configured to perform: performing hidden layer feature extraction on the biometric pattern information to obtain a texture hidden layer feature of the first object, the texture hidden layer feature describing a hidden layer representation of target palmprint information; performing feature extraction on the texture hidden layer feature by a first feature extraction method to obtain the global pattern feature of the first object, the first feature extraction method including a feature dimensionality reduction-based feature extraction method; and performing feature extraction on the texture hidden layer feature by a second feature extraction method to obtain the local pattern feature of the first object, the second feature extraction method including a feature receptive field increase-based feature extraction method.
9 . The computer device according to claim 8 , wherein the one or more processors are further configured to perform: respectively processing the texture hidden layer feature by n second feature extraction methods to obtain n local sub-features of the first object, n being a positive integer, and feature receptive fields of the n second feature extraction methods being different; and splicing the n local sub-features to obtain the local pattern feature of the first object.
10 . The computer device according to claim 7 , wherein the one or more processors are further configured to perform: performing self-attention mechanism processing on the local pattern feature to obtain an updated local pattern feature; and fusing the global pattern feature and the updated local pattern feature of the first object to obtain the fused pattern feature of the first object.
11 . The computer device according to claim 10 , wherein the self-attention mechanism processing comprises an activation process and a regularization process; and the one or more processors are further configured to perform: activating the local pattern feature to obtain a second activated feature; regularizing the local pattern feature to obtain a second regularized feature; and multiplying the second activated feature by the second regularized feature to obtain the updated local pattern feature of the first object.
12 . The computer device according to claim 7 , wherein the one or more processors are further configured to perform: performing interpolation for the biometric pattern information.
13 . A non-transitory computer readable storage medium, storing executable instructions that, when being executed, causes one or more processors implement an identity recognition method by performing: acquiring biometric pattern information describing a biometric pattern of a first object, the biometric pattern information including palm pattern image information; obtaining a plurality of finger gap key points based on the biometric pattern information, wherein the plurality of finger gap key points comprise a first key point, a second key point, and a third key point, the first key point is between an index finger and a middle finger of a palm, the second key point is between the middle finger and a ring finger of the palm, and the third key point is between the ring finger and a little finger of the palm; establishing a rectangular coordinate system in the biometric pattern information, wherein a direction from the third key point to the first key point is a positive direction of an x-axis of the rectangular coordinate system, a direction from an intersection point where a line perpendicular to the x-axis passing through the second key point intersects the x-axis to the second key point as a positive direction of a y-axis of the rectangular coordinate system; determining a central point on the y-axis in a negative y-axis direction, wherein the central point is located on the negative y-axis direction at a particular ratio of a distance from the first key point to the third key point; determining a region of interest in the biometric pattern information based on the central point, the central point being at a center of the region of interest; extracting information in the region of interest to obtain updated biometric pattern information; performing feature extraction on the updated biometric pattern information to obtain a global pattern feature and a local pattern feature, the global pattern feature describing a global feature of the updated biometric pattern information of the first object, and the local pattern feature describing a local feature of the updated biometric pattern information of the first object; fusing the global pattern feature and the local pattern feature of the first object to obtain a fused pattern feature of the first object; and performing identity recognition on the first object based on the fused pattern feature of the first object by comparing the fused pattern feature of the first object with a sample pattern feature to obtain an identity determination result of the first object, the sample pattern feature being used for describing a biometric pattern information feature of a target object, wherein comparing the fused pattern feature of the first object with the sample pattern feature comprises: calculating a cosine similarity between the fused pattern feature and the sample pattern feature; obtaining, in response to the cosine similarity satisfying an identity determination threshold, the identity determination result that the first object is the target object; and obtaining, in response to the cosine similarity not satisfying the identity determination threshold, the identity determination result that the first object is not the target object.
14 . The storage medium according to claim 13 , wherein the one or more processors are further configured to perform: performing hidden layer feature extraction on the biometric pattern information to obtain a texture hidden layer feature of the first object, the texture hidden layer feature describing a hidden layer representation of target palmprint information; performing feature extraction on the texture hidden layer feature by a first feature extraction method to obtain the global pattern feature of the first object, the first feature extraction method including a feature dimensionality reduction-based feature extraction method; and performing feature extraction on the texture hidden layer feature by a second feature extraction method to obtain the local pattern feature of the first object, the second feature extraction method including a feature receptive field increase-based feature extraction method.
15 . The storage medium according to claim 14 , wherein the one or more processors are further configured to perform: respectively processing the texture hidden layer feature by n second feature extraction methods to obtain n local sub-features of the first object, n being a positive integer, and feature receptive fields of the n second feature extraction methods being different; and splicing the n local sub-features to obtain the local pattern feature of the first object.
16 . The storage medium according to claim 13 , wherein the one or more processors are further configured to perform: performing self-attention mechanism processing on the local pattern feature to obtain an updated local pattern feature; and fusing the global pattern feature and the updated local pattern feature of the first object to obtain the fused pattern feature of the first object.

Description

CROSS-REFERENCES TO RELATED APPLICATIONS This application is a continuation application of PCT Patent Application No. PCT/CN2022/132147, filed on Nov. 16, 2022, which claims priority to Chinese Patent Application No. 202210122885.7, on Feb. 9, 2022, all of which is incorporated herein by reference in their entirety. FIELD OF THE TECHNOLOGY The present disclosure relates to the field of identity recognition technologies and, in particular, to an identity recognition method and apparatus, a feature extraction method and apparatus for biometric pattern information, a device, and a storage medium. BACKGROUND OF THE DISCLOSURE With the development of computer technology, application scenarios of biometric pattern information recognition technology continue to expand and is widely used in attendance records, mobile payment, and other scenarios. Biometric pattern information is often recognized by extracting deep feature of the biometric pattern information. During feature extraction, dimensionality reduction is performed on the high-dimensional biometric pattern information to obtain a low-dimensional biometric pattern feature for describing the biometric pattern information. After a comparison process on the biometric pattern features, biometric pattern information recognition is realized. SUMMARY According to an embodiment of the present disclosure, an identity recognition method is provided. The method includes: acquiring biometric pattern information describing a biometric pattern of a first object; performing feature extraction on the biometric pattern information to obtain a global pattern feature and a local pattern feature, the global pattern feature describing a global feature of the biometric pattern information of the first object, and the local pattern feature describing a local feature of the biometric pattern information of the first object; fusing the global pattern feature and the local pattern feature of the first object to obtain a fused pattern feature of the first object; and performing identity recognition on the first object based on the fused pattern feature of the first object. According to an embodiment of the present disclosure, a computer device is provided. The computer device includes: one or more processors and a memory, the memory storing at least one program, and the one or more processor being configured to execute the at least one program to implement an identity recognition method by performing: acquiring biometric pattern information describing a biometric pattern of a first object; performing feature extraction on the biometric pattern information to obtain a global pattern feature and a local pattern feature, the global pattern feature describing a global feature of the biometric pattern information of the first object, and the local pattern feature describing a local feature of the biometric pattern information of the first object; fusing the global pattern feature and the local pattern feature of the first object to obtain a fused pattern feature of the first object; and performing identity recognition on the first object based on the fused pattern feature of the first object. According to an embodiment of the present disclosure, a non-transitory computer-readable storage medium is provided and stores executable instructions that, when being executed, causes one or more processors to implement an identity recognition method by performing: acquiring biometric pattern information describing a biometric pattern of a first object; performing feature extraction on the biometric pattern information to obtain a global pattern feature and a local pattern feature, the global pattern feature describing a global feature of the biometric pattern information of the first object, and the local pattern feature describing a local feature of the biometric pattern information of the first object; fusing the global pattern feature and the local pattern feature of the first object to obtain a fused pattern feature of the first object; and performing identity recognition on the first object based on the fused pattern feature of the first object. BRIEF DESCRIPTION OF THE DRAWINGS In order to describe the technical solutions of the embodiments of the present disclosure more clearly, the following briefly introduces the accompanying drawings required for describing the embodiments. Apparently, the accompanying drawings in the following description show only some embodiments of the present disclosure, and those of ordinary skill in the art may still derive other drawings from these accompanying drawings without involving any creative effort. FIG. 1 is a schematic diagram of a computer system according to an exemplary embodiment of the present disclosure. FIG. 2 is a flowchart of a feature extraction method for biometric pattern information according to an exemplary embodiment of the present disclosure. FIG. 3 is a flowchart of a feature extraction method for biometric pattern information according