A privacy-preserving binary biometric recognition method

Abstract

A method for identifying binary biometric features that can be used for privacy protection, which relates to the technical field of biometric feature identification and information encryption methods. It includes the following steps: first, randomly select some binary bits; randomly divide the binary bits selected by the binary system into two parts, and count the binary bits in the two parts respectively, increase the binary bits corresponding to one part by one, and decrease the binary bits corresponding to the other part by one; repeat step 1 and step 2 for a certain number of times, each binary bit is converted into an integer after counting, and then the newly generated integer feature is stored in the database, and the original feature is deleted. After adopting the above technical solution, the beneficial effect of the present invention is: a biometric feature encoding and identification method for integer encoding storage and identification of binary biometric features, and at the same time, the encoded features can be directly identified, the integer encoding of the method occupies less storage, and the identification calculation process is simple.

Description

Technical Field

[0001] The present invention relates to the technical field of biometric feature recognition and information encryption methods, and in particular to a binary biometric feature recognition method that can be used for privacy protection. Background Art

[0002] A biometric recognition system based on privacy protection refers to a system that effectively identifies biometric features while protecting them. The main difficulty of this system is that it is necessary to encrypt and protect biometric features while ensuring that the protected features can be effectively identified. Currently, security biometric recognition technologies are mainly divided into the following categories:

[0003] Generally, those secure binary biometric identification schemes have been divided into three categories. The first is biometric cryptography. This technology uses digital keys or other mathematical transformations to transform biometric data and enhance security. In order to solve the problem that the biometric data of the same individual cannot be completely consistent, error-correcting codes are usually used to map two biometric codes that differ only in a small part to the same set of codewords. Biometric cryptography systems are further divided into key binding cryptography systems and key generation cryptography systems. The biometrics protected by this method cannot be used directly for identification. The second is revocable biometric technology. The biometric data is transformed into another domain, which enables comparison between different biometric data, and it is crucial that it is difficult to recover the original biometric data from the transformed data. The most representative of these is homomorphic encryption technology. The main problem of this technology is that the transformation process is complicated, which reduces the computational efficiency. The third is negative database technology. This method generates a new data for the biometric. The new data can be used directly for identification, and the original biometric cannot be obtained from the new data. The main problem of this method is that the storage space occupied by the newly generated data is much larger than the original, and the distance calculation during comparison is not concise enough. Summary of the invention

[0004] The purpose of the present invention is to address the defects and shortcomings of the prior art and provide a binary biometric identification method that can be used for privacy protection, which is used to store and identify binary biometrics by integer encoding, and its original binary features cannot be restored from the integer encoding. At the same time, the encoding can be directly similar to the binary biometrics for identification. The binary biometrics can be integer encoded to protect its privacy, and the encoded features can be directly identified by the biometric encoding and identification method. The integer encoding of this method occupies less storage, and the identification calculation process is simple.

[0005] To achieve the above object, the present invention adopts the following technical solution: it comprises the following steps:

[0006] Step 1: First, randomly select some binary bits;

[0007] Step 2: randomly divide the binary digits selected by the binary system into two parts, and count the binary digits in the two parts respectively, increase the binary digits corresponding to one part by one, and decrease the binary digits corresponding to the other part by one;

[0008] Step 3: Repeat steps 1 and 2 for a certain number of times. Each binary bit is converted into an integer after counting. Then the newly generated integer feature is stored in the database and the original feature is deleted.

[0009] Step 4: When performing a comparison, the binary biometric feature can be directly compared with the integer code to obtain a similarity measure, thereby identifying the biometric feature.

[0010] The specific steps of integer encoding of the binary biometric feature are as follows:

[0011] Step 1: The encoding module converts the biometric feature B into an n-bit binary code (b1,...,b n ), and set a zero vector v=(v0,...,v n );

[0012] Step 2: Set a sufficiently large number M, and set m=0;

[0013] Step 3: If m < M, go to step 4; otherwise go to step 10;

[0014] Step 4: Randomly select L positions from the n positions to form a set S = {i1,...,i L ,j1,...,j L}, where i Y ,j Y All correspond to the y-th binary bit;

[0015] Step 5: Generate a new set S' = {a k |a k is a subset of S and a k does not contain i at the same time h and j h , 1≤h≤L};

[0016] Step 6: Select q subsets of the set S' and design a probability p1,...,p for each subset q , and p1+...+p q =1;

[0017] Step 7, generate a random number 0<α<1, and find z such that Where p0=0;

[0018] Step 8. Select set a z ∈S', let (V i =V i -1 and V j =V j +1);

[0019] Step 9, let m=m+1, and execute step 3;

[0020] Step 10: Perform the following operations on each dimension of V: ​​If the bit b w The value of is 0, then V w The value becomes V w The opposite of

[0021] Step 11: Store V as the encoded data of B into the biometric database, and then delete B.

[0022] The specific steps of comparing the binary biometric feature with the integer code are as follows:

[0023] Step 1: Assume the current biometric feature is B, and convert B into an n-bit binary code (b1,...,b n ), assuming i=1;

[0024] Step 2: Assume that the number of data entries stored in the biometric database is m. If i≤m, execute step 3; otherwise, execute step;

[0025] Step 3: Let the i-th record in the database be V i , let C=(c1,...,c n )=-2×(b1-1,...,b n -1), calculate

[0026] Step 4: Let i=i+1, and execute step 2;

[0027] Step 5: In d1,...,d m Select the smallest value d min , compare it with the set threshold, if it is less than the threshold, the biometric matches, otherwise it does not match.

[0028] In the comparison between the binary biometric feature and the integer code, some bits may be masked, and the specific steps are as follows:

[0029] Step 1: Set the n-dimensional binary vector mask = (h1,...,h n ), where the mask bit is set to 0 and the non-mask bit is set to 1. Let the current biometric feature be B, and convert B into an n-bit binary code (b1,...,b n ), assuming i=1;

[0030] Step 2: Assume that the number of data entries stored in the biometric database is m. If i≤m, execute step 3; otherwise, execute step;

[0031] Step 3: Let the i-th record in the database be V i , let C=(c1,...,c n )=-2×(b1-1,...,b n -1), calculate

[0032] Step 4: Let i=i+1, and execute step 2;

[0033] Step 5: In d1,...,d m Select the smallest value d min , compare it with the set threshold, if it is less than the threshold, the biometric matches, otherwise it does not match.

[0034] A shift operation can also be performed in the comparison between the binary biometric feature and the integer code, and the specific steps are as follows:

[0035] Step 1: Assume the current biometric feature is B, and convert B into an n-bit binary code (b1,...b n ), let i = 1, shift the number of bits to k;

[0036] Step 2: Assume that the number of data entries stored in the biometric database is m. If i≤m, execute step 3; otherwise, execute step;

[0037] Step 3: Let the i-th record in the database be V i , let C=(c1,...,c n )=-2×(b1-1,...,b n -1), calculate

[0038] Step 4: Let i=i+1, and execute step 2;

[0039] Step 5: In d1,...,d m Select the smallest value d min , compare it with the set threshold, if it is less than the threshold, the biometric matches, otherwise it does not match.

[0040] The working principle of the present invention is as follows: first, a part of binary bits are randomly selected; the binary bits selected in the binary system are randomly divided into two parts, and the binary bits in the two parts are counted respectively, the binary bits corresponding to one part are increased by one, and the binary bits corresponding to the other part are decreased by one; steps 1 and 2 are repeated for a certain number of times, each binary bit is converted into an integer after counting, and then the newly generated integer feature is stored in the database, and the original feature is deleted; when comparing, the binary biometric feature can be directly compared with the integer code, and the similarity measurement can be obtained, so as to identify the biometric feature.

[0041] After adopting the above technical scheme, the beneficial effects of the present invention are: it is used to perform integer encoding, storage and recognition of binary biometric features, and its original binary features cannot be restored from the integer encoding. At the same time, the encoding can be directly calculated with the binary biometric features for similarity calculation for recognition. It can not only perform integer encoding on binary biometric features to protect their privacy, but also directly recognize the encoded features. The integer encoding of this method occupies less storage, and the recognition calculation process is simple. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0043] Figure 1 It is a schematic block diagram of the process of the present invention;

[0044] Figure 2 It is a schematic block diagram of the process of integer encoding of binary biometric features in the present invention;

[0045] Figure 3 is a schematic flow chart of the comparison between binary biometric features and integer codes in the present invention;

[0046] Figure 4 It is a schematic flow chart of the mid-level shielding operation of the present invention;

[0047] Figure 5 It is a schematic flow chart of the shift operation in the present invention. DETAILED DESCRIPTION

[0048] See also Figure 1-Figure 5 As shown, the technical solution adopted in this specific implementation is: it comprises the following steps:

[0049] Step 1: First, randomly select some binary bits;

[0050] Step 2: randomly divide the binary digits selected by the binary system into two parts, and count the binary digits in the two parts respectively, increase the binary digits corresponding to one part by one, and decrease the binary digits corresponding to the other part by one;

[0051] Step 3: Repeat steps 1 and 2 for a certain number of times. Each binary bit is converted into an integer after counting. Then the newly generated integer feature is stored in the database and the original feature is deleted.

[0052] Step 4: When performing a comparison, the binary biometric feature can be directly compared with the integer code to obtain a similarity measure, thereby identifying the biometric feature.

[0053] The specific steps of integer encoding of the binary biometric feature are as follows:

[0054] Step 1: The encoding module converts the biometric feature B into an n-bit binary code (b1,...,b n ), and set a zero vector v=(v0,...,v n );

[0055] Step 2: Set a sufficiently large number M, and set m=0;

[0056] Step 3: If m < M, go to step 4; otherwise go to step 10;

[0057] Step 4: Randomly select L positions from the n positions to form a set S = {i1,...,i L ,j1,...,j L}, where i Y ,j Y All correspond to the y-th binary bit;

[0058] Step 5: Generate a new set S' = {a k |a k is a subset of S and a k does not contain i at the same time h and j h , 1≤h≤L};

[0059] Step 6: Select q subsets of the set S' and design a probability p1,...,p for each subset q , and p1+...+p q =1;

[0060] Step 7, generate a random number 0<α<1, and find z such that Where p0=0;

[0061] Step 8. Select set az ∈S', let (V i =V i -1 and V j =V j +1);

[0062] Step 9, let m=m+1, and execute step 3;

[0063] Step 10: Perform the following operations on each dimension of V: ​​If the bit b w The value of is 0, then V w The value becomes V w The opposite of

[0064] Step 11: Store V as the encoded data of B into the biometric database, and then delete B.

[0065] The specific steps of comparing the binary biometric feature with the integer code are as follows:

[0066] Step 1: Assume the current biometric feature is B, and convert B into an n-bit binary code (b1,...,b n ), assuming i=1;

[0067] Step 2: Assume that the number of data entries stored in the biometric database is m. If i≤m, execute step 3; otherwise, execute step;

[0068] Step 3: Let the i-th record in the database be V i , let C=(c1,...,c n )=-2×(b1-1,...,b n -1), calculate

[0069] Step 4: Let i=i+1, and execute step 2;

[0070] Step 5: In d1,...,d m Select the smallest value d min , compare it with the set threshold, if it is less than the threshold, the biometric matches, otherwise it does not match.

[0071] In the comparison between the binary biometric feature and the integer code, some bits may be masked, and the specific steps are as follows:

[0072] Step 1: Set the n-dimensional binary vector mask = (h1,...,h n ), where the mask bit is set to 0 and the non-mask bit is set to 1. Let the current biometric feature be B, and convert B into an n-bit binary code (b1,...,b n ), assuming i=1;

[0073] Step 2: Assume that the number of data entries stored in the biometric database is m. If i≤m, execute step 3; otherwise, execute step;

[0074] Step 3: Let the i-th record in the database be V i , let C=(c1,...,c n )=-2×(b1-1,...,b n -1), calculate

[0075] Step 4: Let i=i+1, and execute step 2;

[0076] Step 5: In d1,...,d m Select the smallest value d min , compare it with the set threshold, if it is less than the threshold, the biometric matches, otherwise it does not match.

[0077] A shift operation can also be performed in the comparison between the binary biometric feature and the integer code, and the specific steps are as follows:

[0078] Step 1: Assume the current biometric feature is B, and convert B into an n-bit binary code (b1,...b n ), let i = 1, shift the number of bits to k;

[0079] Step 2: Assume that the number of data entries stored in the biometric database is m. If i≤m, execute step 3; otherwise, execute step;

[0080] Step 3: Let the i-th record in the database be V i , let C=(c1,...,c n )=-2×(b1-1,...,b n -1), calculate

[0081] Step 4: Let i=i+1, and execute step 2;

[0082] Step 5: In d1,...,d m Select the smallest value d min , compare it with the set threshold, if it is less than the threshold, the biometric matches, otherwise it does not match.

[0083] The working principle of the present invention is as follows: first, a part of binary bits are randomly selected; the binary bits selected in the binary system are randomly divided into two parts, and the binary bits in the two parts are counted respectively, the binary bits corresponding to one part are increased by one, and the binary bits corresponding to the other part are decreased by one; steps 1 and 2 are repeated for a certain number of times, each binary bit is converted into an integer after counting, and then the newly generated integer feature is stored in the database, and the original feature is deleted; when comparing, the binary biometric feature can be directly compared with the integer code, and the similarity measurement can be obtained, so as to identify the biometric feature.

[0084] After adopting the above technical scheme, the beneficial effects of the present invention are: it is used to perform integer encoding, storage and recognition of binary biometric features, and its original binary features cannot be restored from the integer encoding. At the same time, the encoding can be directly calculated with the binary biometric features for similarity calculation for recognition. It can not only perform integer encoding on binary biometric features to protect their privacy, but also directly recognize the encoded features. The integer encoding of this method occupies less storage, and the recognition calculation process is simple.

[0085] The above description is only used to illustrate the technical solution of the present invention rather than to limit it. Other modifications or equivalent substitutions made to the technical solution of the present invention by ordinary technicians in this field should be included in the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.

Claims

1. A binary biometric recognition method for privacy protection, characterized in that: It includes the following steps: Step 1: First, randomly select some binary bits; Step 2: randomly divide the binary digits selected by the binary system into two parts, and count the binary digits in the two parts respectively, increase the binary digits corresponding to one part by one, and decrease the binary digits corresponding to the other part by one; Step 3: Repeat steps 1 and 2 for a certain number of times. Each binary bit is converted into an integer after counting. Then the newly generated integer feature is stored in the database and the original feature is deleted. Step 4: When performing the comparison, the binary biometric feature can be directly compared with the integer code to obtain the similarity measure, thereby identifying the biometric feature. The specific steps of comparing the binary biometric feature with the integer code are as follows: Step 1: Assume the current biometric feature is B, and convert B into an n-bit binary code (b 1,..., b n ), assuming i=1; Step 2: Assume that the number of data entries stored in the biometric database is m. If i≤m, execute step 3; otherwise, execute step; Step 3: Let the i-th record in the database be V i , let C = (c 1,..., c n )=-2×(b1-1,...,b n -1), calculate Step 4: Let i=i+1, and execute step 2; Step 5: 1,..., d m Select the smallest value d min , compare it with the set threshold, if it is less than the threshold, the biometric matches, otherwise it does not match.

2. A method for identifying binary biometric features that can be used for privacy protection according to claim 1, characterized in that: The specific steps of integer encoding of the binary biometric feature are as follows: Step 1: The encoding module converts the biometric feature B into an n-bit binary code (b 1,..., b n ), and set a zero vector v=(v0,...,v n ); Step 2: Set a sufficiently large number M, and set m=0; Step 3: If m<M, go to step 4; Otherwise, go to step 10; Step 4: Randomly select L positions from the n positions to form a set S = {i1,...,i L ,j1,...,j L }, where i Y ,j Y All correspond to the y-th binary bit; Step 5: Generate a new set S' = {a k |a k is a subset of S and a k does not contain i at the same time h and j h , 1≤h≤L}; Step 6: Select q subsets of the set S' and design a probability p1,...,p for each subset q , and p1+...+p q =1; Step 7, generate a random number 0<α<1, and find z such that Where p0=0; Step 8. Select set a z ∈S', let (V i =V i -1 and V j =V j +1); Step 9, let m=m+1, and execute step 3; Step 10: Perform the following operations on each dimension of V: ​​If the bit b w The value of is 0, then V w The value becomes V w The opposite of Step 11: Store V as the encoded data of B into the biometric database, and then delete B.

3. A method for binary biometric identification that can be used for privacy protection according to claim 1, characterized in that: In the comparison between the binary biometric feature and the integer code, some bits may be masked, and the specific steps are as follows: Step 1: Set the n-dimensional binary vector mask = (h1,...,h n ), where the mask bit is set to 0 and the non-mask bit is set to 1. Let the current biometric feature be B, and convert B into an n-bit binary code (b1,...,b n ), assuming i=1; Step 2: Assume that the number of data entries stored in the biometric database is m. If i≤m, execute step 3; otherwise, execute step; Step 3: Let the i-th record in the database be V i , let C=(c1,...,c n )=-2×(b1-1,...,b n -1), calculate Step 4: Let i=i+1, and execute step 2; Step 5: 1,..., d m Select the smallest value d min , compare it with the set threshold, if it is less than the threshold, the biometric matches, otherwise it does not match.

4. A method for binary biometric identification that can be used for privacy protection according to claim 1, characterized in that: A shift operation can also be performed in the comparison between the binary biometric feature and the integer code, and the specific steps are as follows: Step 1: Assume the current biometric feature is B, and convert B into an n-bit binary code (b1,...b n ), let i = 1, shift the number of bits to k; Step 2: Assume that the number of data entries stored in the biometric database is m. If i≤m, execute step 3; otherwise, execute step; Step 3: Let the i-th record in the database be V i , let C=(c1,...,c n )=-2×(b1-1,...,b n -1), calculate Step 4: Let i=i+1, and execute step 2; Step 5: In d1,...,d m Select the smallest value d min , compare it with the set threshold, if it is less than the threshold, the biometric matches, otherwise it does not match.

Images