Fingerprint registration and verification method and system for providing differential privacy protection

Abstract

The invention relates to a fingerprint registration and verification method and system for providing differential privacy protection, and belongs to the technical field of biological feature recognition. According to the method and the device, the fingerprint data feature transformation based on the differential privacy protection technology is realized, so that the function that the original fingerprint data of the user does not leave the local but still can be subjected to fingerprint registration and verification is realized, and the aim of ensuring that the privacy protection of the fingerprint data of the user and the data availability are both considered is fulfilled.

Description

Technical Field

[0001] The present invention belongs to the technical field of biometric identification, and in particular relates to a fingerprint registration and verification method and system providing differential privacy protection. Background Art

[0002] Biometric identification technology refers to the technology of automatic identity authentication based on the physiological or behavioral characteristics inherent in the human body. Fingerprint is one of the most commonly used physiological characteristics in the identity authentication mechanism. Generally speaking, when completing the fingerprint data identity authentication process, users need to upload the original fingerprint data to the server that stores and verifies the information, which poses a risk of data privacy leakage. Summary of the invention

[0003] 1. Technical issues to be resolved

[0004] The technical problem to be solved by the present invention is to design a fingerprint registration and method that provides differential privacy protection, so as to realize the function of fingerprint registration and verification without leaving the local area of ​​the user's original fingerprint data, and ensure both the privacy protection of the user's fingerprint data and the availability of the data.

[0005] (II) Technical solution

[0006] In order to solve the above technical problems, the present invention provides a fingerprint registration and verification method providing differential privacy protection, comprising the following steps:

[0007] S101, when a user registers a fingerprint, a local fingerprint collection device is used to obtain the original fingerprint data of the user to obtain a d-dimensional fingerprint feature vector T;

[0008] S102, performing feature transformation to provide differential privacy protection on the fingerprint feature vector T required in the fingerprint registration process to obtain a transformed fingerprint feature vector P1, repeating the process N times to obtain a transformation set P = {P1, P2, ..., PN} of the fingerprint feature vector T, where N is an odd number;

[0009] S103, sending the transformation set P of the fingerprint feature vector T to the storage and verification server, so that the user completes the fingerprint registration through the transformation set P received by the storage and verification server;

[0010] S104, when the user performs fingerprint verification, the local fingerprint collection device is used to obtain the original fingerprint data of the user, and the fingerprint feature vector T' is extracted;

[0011] S105, performing a feature transformation to provide differential privacy protection on the fingerprint feature vector T' required for the fingerprint verification process, and obtaining a transformed fingerprint feature vector Pq; wherein the same process as that for calculating P1 in S102 is adopted to obtain a transformed set Pq of the fingerprint feature vector;

[0012] S106, sending the transformation set Pq to the storage and verification server, performing a similarity comparison between the elements in Pq and the elements in the transformation set P one by one, and then performing a majority vote on the comparison results to obtain a fingerprint verification result, so that the user completes the fingerprint verification.

[0013] The present invention also provides a system for implementing the method.

[0014] (III) Beneficial effects

[0015] The present invention provides a fingerprint registration and verification method and system that provide differential privacy protection. By implementing fingerprint data feature transformation based on differential privacy protection technology, the user's original fingerprint data does not leave the local but can still be used for fingerprint registration and verification, thereby achieving the purpose of ensuring both user fingerprint data privacy protection and data availability. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a flow chart of a fingerprint registration and verification method providing differential privacy protection provided by an embodiment of the present invention;

[0017] Figure 2 A schematic diagram of the structure of a fingerprint registration and verification system providing differential privacy protection provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, content and advantages of the present invention more clear, the specific implementation methods of the present invention are further described in detail below in conjunction with the drawings and examples.

[0019] The present invention provides a fingerprint registration and verification method for providing differential privacy protection, and the overall process includes: when a user performs fingerprint registration, the original fingerprint data of the user is obtained by using a local fingerprint acquisition device, and a fingerprint feature vector T is extracted; the fingerprint feature vector T corresponding to the fingerprint registration is subjected to a feature transformation for providing differential privacy protection to obtain a transformed fingerprint feature vector P1, and the process is repeated N times to obtain a transformation set P={P1, P2, ..., PN} of the fingerprint feature vector; the transformation set P of the fingerprint feature vector is sent to a storage and verification server, so that the user completes the fingerprint registration; in response to the user performing fingerprint verification, the original fingerprint data of the user is obtained by using a local fingerprint acquisition device, and a fingerprint feature vector T' is extracted; the fingerprint feature vector T' corresponding to the fingerprint verification is subjected to a feature transformation for providing differential privacy protection to obtain a transformed fingerprint feature vector Pq; the transformed fingerprint feature vector Pq is sent to the storage and verification server, and the elements in the transformation set P of the fingerprint feature vector are compared for similarity one by one, and then the comparison results are subjected to majority voting to obtain a verification result, so that the user completes the fingerprint verification.

[0020] Based on the above design ideas, the embodiment of the present invention provides a fingerprint registration and verification method that provides differential privacy protection, see Figure 1 , the method comprises the following steps:

[0021] S101, when a user registers a fingerprint, a local fingerprint collection device is used to obtain the original fingerprint data of the user to obtain a d-dimensional fingerprint feature vector T;

[0022] S102, perform feature transformation to provide differential privacy protection on the fingerprint feature vector T required in the fingerprint registration process to obtain the transformed fingerprint feature vector P1, repeat the process N times to obtain the transformation set P = {P1, P2, ..., PN} of the fingerprint feature vector T, where N is an odd number; the details are as follows:

[0023] First, for the d-dimensional fingerprint feature vector T = {t1, t2, ..., td}, N groups of hash functions are preset, each group contains K independent hash functions, and a window of length W is preset, W ≤ d, and the privacy budget ε in differential privacy protection;

[0024] Then, select the kth function among the K hash functions in the nth group and apply it to the fingerprint feature vector T to reorder the elements in T and obtain the updated fingerprint feature vector hk(T), n=1,...,,N, with an initial value of 1; then use a window of length W to align hk(T) from the left and retain the elements in the window (the elements in the window are the elements in hk(T)) to obtain the shortened fingerprint feature vector hk(T,W); at this time, return the ranking value rk of the element with the largest value in the vector hk(T,W), where the ranking values ​​are accumulated from left to right; then use the Laplace mechanism to add noise to the ranking value rk and round it up to obtain the ranking value rk' of differential privacy protection that provides a privacy budget ε (the parameter ε is used in the algorithm that uses the Laplace mechanism to implement differential privacy protection). After K cycles of the above process, P1={r1',r2',...,rK'} is obtained.

[0025] Among them, rK' is the result after adding noise to rK through the Laplace mechanism. rK is generated based on the Kth hash function in the nth group, and the generation process is:

[0026] 1) Under the action of the Kth hash function, a vector hK(T) is obtained from the vector T (the Kth function among the K hash functions in the nth group is selected and applied to the fingerprint feature vector T to reorder the elements in T and obtain the updated fingerprint feature vector hK(T));

[0027] 2) Then, under the action of a window of length W, a shortened vector hK(T,W) is obtained;

[0028] 3) Return the ranking value of the largest element in the vector hK(T,W), which is rK. For example, assuming that the vector hK(T,W) = {t3, t9, t2, t7}, where t3 = 15, t9 = 10, t2 = 20, t7 = 16, then the largest element in the vector is t2, and its ranking value is 3, which means rK = 3.

[0029] Finally, the hash function of the n+1th group is selected to repeat the above process to obtain P = {P1, P2, ..., PN}.

[0030] S103, sending the transformation set P of the fingerprint feature vector T to the storage and verification server, so that the user completes the fingerprint registration through the transformation set P received by the storage and verification server;

[0031] S104, when the user performs fingerprint verification, the local fingerprint collection device is used to obtain the original fingerprint data of the user, and the fingerprint feature vector T' is extracted;

[0032] S105, performing feature transformation on the fingerprint feature vector T' required for the fingerprint verification process to provide differential privacy protection, to obtain a transformed fingerprint feature vector Pq;

[0033] The same process as that for calculating P1 in S102 is adopted to obtain the transformation set Pq={rq1, rq2,…, rqK} of the fingerprint feature vector.

[0034] S106, sending the transformation set Pq to the storage and verification server, performing a similarity comparison between the elements in Pq and the elements in the transformation set P one by one, and then performing a majority vote on the comparison results to obtain a fingerprint verification result, so that the user completes the fingerprint verification.

[0035] Specifically, the PO (Pairwise Order) similarity between Pq and the elements in P is calculated to obtain N comparison results, where if the similarity exceeds a preset threshold, it is considered that the comparison is passed, and a majority vote is performed on the N comparison results to obtain the final fingerprint verification result. Since N is set to an odd number in step S102, a majority vote in this step to obtain the final result will definitely result in a result (if it is an even number, it is possible that half of the comparisons pass and half fail, and no result can be voted).

[0036] See also Figure 2 In an embodiment of the present invention, a fingerprint registration and verification system providing differential privacy protection is also provided, the system comprising:

[0037] The data acquisition unit 10 is used to obtain the original fingerprint data of the user using a local fingerprint acquisition device and extract the fingerprint feature vector when the user performs fingerprint registration or verification;

[0038] A feature transformation unit 20, configured to perform a feature transformation for providing differential privacy protection on a fingerprint feature vector corresponding to the fingerprint registration or verification, to obtain a transformed fingerprint feature vector;

[0039] A data storage unit 30, used to store the transformed fingerprint feature vector;

[0040] The data comparison unit 40 is used to perform similarity comparison on the transformed fingerprint feature vectors stored in the data storage unit during registration and verification.

[0041] It can be seen that the present invention realizes the fingerprint data feature transformation based on differential privacy protection technology, and further realizes the function of fingerprint registration and verification without leaving the user's original fingerprint data, ensuring both the privacy protection of user fingerprint data and data availability.

[0042] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A fingerprint registration and verification method providing differential privacy protection, characterized in that: The following steps are involved: S101, when a user registers a fingerprint, a local fingerprint collection device is used to obtain the original fingerprint data of the user to obtain a d-dimensional fingerprint feature vector T; S102, performing feature transformation to provide differential privacy protection on the fingerprint feature vector T required in the fingerprint registration process to obtain a transformed fingerprint feature vector P1, repeating the process N times to obtain a transformation set P = {P1, P2, ..., PN} of the fingerprint feature vector T, where N is an odd number; S103, sending the transformation set P of the fingerprint feature vector T to the storage and verification server, so that the user completes the fingerprint registration through the transformation set P received by the storage and verification server; S104, when the user performs fingerprint verification, the local fingerprint collection device is used to obtain the original fingerprint data of the user, and the fingerprint feature vector T' is extracted; S105, performing a feature transformation to provide differential privacy protection on the fingerprint feature vector T' required for the fingerprint verification process, and obtaining a transformed fingerprint feature vector Pq; wherein the same process as that for calculating P1 in S102 is adopted to obtain a transformed set Pq of the fingerprint feature vector; S106, sending the transformation set Pq to the storage and verification server, performing a similarity comparison between the elements in Pq and the elements in the transformation set P one by one, and then performing a majority vote on the comparison results to obtain a fingerprint verification result, so that the user completes the fingerprint verification.

2. The method according to claim 1, characterized in that Step S102 is specifically as follows: First, for the d-dimensional fingerprint feature vector T = {t1, t2, ..., td}, N groups of hash functions are preset, each group contains K independent hash functions, and a window of length W is preset, W ≤ d, and the privacy budget ε in differential privacy protection; Then, select the kth function among the K hash functions in the nth group and apply it to the fingerprint feature vector T to reorder the elements in T to obtain the updated fingerprint feature vector hk(T), n=1,...,,N, and the initial value is 1; then use a window of length W to align with hk(T) from the left and retain the elements in the window (the elements in the window are the elements in hk(T)) to obtain the shortened fingerprint feature vector hk(T,W); at this time, return the sort value rk of the element with the largest value in the vector hk(T,W), where sorting is from left to right, and sorting is The values ​​are accumulated starting from 1; the Laplace mechanism is then used to add noise to the ranking value rk and round it up to get the ranking value rk' that provides differential privacy protection with a privacy budget ε (the parameter ε is used in the algorithm that uses the Laplace mechanism to implement differential privacy protection); after K cycles, P1 = {r1', r2', ..., rK'} is obtained; the ranking value rK' that provides differential privacy protection is the result of adding noise to the ranking value rK through the Laplace mechanism; the ranking value rK is generated based on the Kth hash function in the nth group; Finally, the hash function of the n+1th group is selected to repeat the above process to obtain P = {P1, P2, ..., PN}.

3. The method according to claim 2, characterized in that The generation process of the ranking value rK is: 1) Under the action of the Kth hash function, a vector hK(T) is obtained from the vector T; 2) Then, under the action of a window of length W, a shortened vector hK(T,W) is obtained; 3) Return the sorted value of the element with the largest value in the vector hK(T,W), which is rK.

4. The method according to claim 1, characterized in that In step S106, the PO similarity between Pq and the elements in P is calculated to obtain N comparison results, wherein if the similarity exceeds a preset threshold, it is considered that the comparison passes, and a majority vote is performed on the N comparison results to obtain the final fingerprint verification result.

5. The method according to claim 1, characterized in that Application in the field of biometric recognition technology.

6. The method according to claim 1, characterized in that Application in the field of network security technology.

7. The method according to claim 1, characterized in that Application in the field of computer technology.

8. A system for implementing the method according to any one of claims 1 to 7.

9. The system according to claim 8, characterized in that The system includes: The data collection unit is used to obtain the original fingerprint data of the user using the local fingerprint collection device when the user registers or verifies the fingerprint, and extract the fingerprint feature vector; A feature transformation unit, configured to perform a feature transformation for providing differential privacy protection on a fingerprint feature vector corresponding to the fingerprint registration or verification, to obtain a transformed fingerprint feature vector; A data storage unit, used for storing the transformed fingerprint feature vector; The data comparison unit is used to perform similarity comparison on the transformed fingerprint feature vectors stored in the data storage unit during registration and verification.

10. The system according to claim 8, characterized in that Application in the field of biometric recognition technology.

Images