A bio-information authentication method and device based on a stream cipher, and a medium

Abstract

The application relates to a biological information authentication method and device based on a stream cipher, and a medium, the method is applied to a server side, and comprises the following steps: establishing a connection with a client, obtaining a preprocessed face biological feature vector from the client; obtaining an original password from the client, generating key information according to the original password; encrypting the face biological feature vector and the key information through a 256-stream cipher to obtain encrypted ciphertext; calculating the similarity index of the encrypted ciphertext and preset original biological information, judging whether the similarity index is less than a preset threshold, if yes, authentication is passed, and if not, authentication fails and a prompt information is sent. Compared with the prior art, the 256-stream cipher encryption mode is adopted, and the security and performance are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of information security technology, and in particular to a biometric authentication method, device, and medium based on stream cipher. Background Technology

[0002] Biometric authentication is a technology that identifies people based on their physiological and behavioral characteristics. The selection of biometric features should follow the principles of permanence, uniqueness, universality, robustness, and collectability.

[0003] A highly sensitive issue in biometric authentication is the security of the biometric information itself. Due to its unique characteristics, such as permanence and uniqueness, biometric information has become a popular authentication method. However, these characteristics are a double-edged sword. Permanence means it will never change, ensuring successful authentication every time, but it also means that once leaked, it's permanently lost. Uniqueness means attackers cannot copy it, but it also means it absolutely cannot be lost; if lost, there are no new biometric characteristics to replace it.

[0004] The security protection of biometric information has always been a key research issue in biometric authentication. Security issues in biometric authentication can be broadly categorized into two types: privacy-preserving computation and biometric template protection. Privacy-preserving computation typically requires the server to operate on encrypted data, such as using homomorphic encryption or obfuscated circuits. One possible use case is when the client obtains the original biometric information and needs to hand it over to the server for feature extraction. In this case, the server's feature extraction process is performed on encrypted ciphertext, and the subsequent Euclidean distance comparison is also performed on encrypted feature vectors. This is called privacy-preserving computation, or security in computation. The other type of security refers to biometric template security, which aims to securely store biometric information in a database. This patent studies a protection scheme for biometric template security.

[0005] Research on existing biometric template security mainly falls into two categories: cancelable biometrics and biometric cryptosystems. Cancelable biometric templates refer to transforming the template feature vector using irreversible functions, such as hash functions, and then storing the transformed vector in a database. Even if an attacker steals the feature vector from the database, because this feature vector is already transformed, the one-way nature of the transformation prevents the attacker from obtaining the original biometric vector. Currently, biometric cryptosystems employ two methods: key generation and key binding. Key generation uses a fuzzy extractor to directly generate a key from the original biometric template and then saves the key. This method only leaks an auxiliary string and does not reveal any biometric information because the stored information in the database is not the original biometric information. Key binding is another method, such as FuzzyCommitment or Fuzzy Vault schemes. In this method, the key needs to be selected separately, and error-correcting codewords can be encoded using the key. Auxiliary data is obtained by XORing the codewords with the biometric template. This method also does not leak the original biometric information. Both methods are feasible.

[0006] Chinese patent application CN201210421519.8 discloses an identity authentication method, an identity authentication server, and an identity authentication device. The method includes: an authentication server receiving a biometric authentication request from a client; the authentication server returning a randomly generated challenge number to the client; the authentication server receiving a dynamic verification password from the client, wherein the dynamic verification password is generated by the client's time when the challenge number is received, the challenge number, the user's biometrics, and a first seed key pre-stored by the client; the authentication server generating a dynamic password based on the authentication server's time when the dynamic verification password is received, the challenge number, the user's biometrics, and a second seed key; and the authentication server verifying whether the dynamic password and the dynamic verification password are consistent, and returning an identity authentication result to the client based on the verification result. This application uses biometrics to ensure the security of the dynamic password, thus guaranteeing the authenticity and reliability of user identity authentication. However, this application does not address the problem that traditional cryptographic protection schemes can affect the effectiveness of biometric identification.

[0007] In summary, there is currently a lack of biometric authentication methods to address the issue that traditional cryptographic protection schemes can negatively impact the effectiveness of biometric identification. Summary of the Invention

[0008] The purpose of this invention is to overcome the defects of the prior art by providing a biometric authentication method, device, and medium based on stream cipher, so as to solve or partially solve the problem that traditional cryptographic protection schemes affect the effectiveness of biometric identification.

[0009] The objective of this invention can be achieved through the following technical solutions:

[0010] One aspect of the present invention provides a stream cipher-based biometric authentication method applied to a server, comprising the following steps:

[0011] Establish a connection with the client and obtain a preprocessed facial biometric vector from the client;

[0012] Obtain the original password from the client and generate key information based on the original password;

[0013] Based on the facial biometric vector and the key information, the encrypted text is obtained by encrypting it using a 256-stream cipher.

[0014] For the encrypted ciphertext and the preset original biometric information, calculate the similarity index between the two, and determine whether the similarity index is less than a preset threshold. If yes, the authentication is successful; otherwise, the authentication fails and a prompt message is sent.

[0015] As a preferred technical solution, the process of obtaining the facial biometric vector includes the following steps:

[0016] The facial biometric vector is obtained by sequentially processing the original image, performing face recognition and feature extraction via a deep convolutional neural network.

[0017] As a preferred technical solution, the process of obtaining the key information includes the following steps:

[0018] For the original password, the obfuscated password information is obtained by adding salt, and a pseudo-random number is extracted using a preset random extractor as the key information.

[0019] As a preferred technical solution, the facial biometric vector is a decimal multidimensional vector.

[0020] As a preferred technical solution, the encrypted ciphertext is obtained using the following formula:

[0021] c1 = (m + key1) mod 256

[0022] p = (m + key1) > 255? 1:0

[0023]

[0024] In the formula, m is the value of a certain dimension in the facial biometric feature vector, c1 and s1 constitute the value of the encrypted ciphertext in the corresponding dimension, key1 is the generated first key, mod means modulo, when (m+key1)>255, p=1, when (m+key1)≤255, p=0, and key2∈(0,1) is used to cover one bit of p.

[0025] As a preferred technical solution, the similarity index is Euclidean distance.

[0026] As a preferred technical solution, for the encrypted ciphertext of the form (c1, s1) and the original biometric information of the form (c2, s2), the acquisition of the similarity index includes the following steps:

[0027] judge If the similarity index is equal to 0, then the similarity index is |E(c1)-E(c2)|; otherwise, the similarity index is 256-|E(c1)-E(c2)|.

[0028] As a preferred technical solution, the facial biometric feature vector takes the value of an integer between 0 and 255 in each dimension.

[0029] In another aspect, an electronic device is provided, comprising: one or more processors and a memory, wherein the memory stores one or more programs, the one or more programs including instructions for executing the above-described stream cipher-based biometric authentication method.

[0030] In another aspect, the present invention provides a computer-readable storage medium including one or more programs executable by one or more processors of an electronic device, said one or more programs including instructions for performing the above-described stream cipher-based biometric authentication method.

[0031] Compared with the prior art, the present invention has the following advantages:

[0032] (1) This invention possesses the basic characteristics of irreversibility, updability and non-linkability in biometrics. Since it uses 256-stream cipher for encryption, the similarity index will not change. Compared with traditional cryptographic protection schemes, which lead to reduced distinguishability and entropy loss after various conversions or encryptions, the proposed scheme satisfies that the recognition accuracy of the converted biometric template is equal to that of the original biometric template. This invention protects biometric information without losing recognition accuracy, achieving the effect of unchanged recognition performance before and after encryption, thereby solving or partially solving the problem that traditional cryptographic protection schemes affect the biometric recognition effect.

[0033] (2) This invention does not require additional number system conversion, which effectively simplifies the authentication process and allows direct encryption of facial biometric vectors without feature conversion.

[0034] (3) Traditional multimodal biometric authentication schemes mostly require the use of a fuzz extractor. However, this component has some drawbacks. For example, the auxiliary data generated during the registration process may leak the link information of the original biometric template. In contrast, the present invention does not require the use of a fuzz extractor. Attached Figure Description

[0035] Figure 1 This is an architecture diagram of the biometric authentication scheme based on stream cipher in Example 1;

[0036] Figure 2 A schematic diagram illustrating the process of facial information extraction and transformation;

[0037] Figure 3 This is a schematic diagram of the various functional modules and their input / output relationships in Example 1. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0039] Example 1

[0040] like Figure 1-3 This embodiment provides a biometric authentication method based on stream ciphers, including the following steps:

[0041] Step S1, Feature Extraction. Feature extraction uses ArcSoft's ArcFace SDK for face recognition. The acquired face image is input into the ArcFace SDK via a developed Android application to extract feature vectors. The core of using a Deep Convolutional Neural Network (DCNN) for biometric authentication lies in designing a suitable loss function to enhance judgment capabilities. Training the DCNN for face recognition primarily involves training the multi-class classifier Softmax. ArcFace's face processing flow includes preprocessing, training, and face testing, ultimately generating a decimal integer biometric feature vector. The output biometric feature vector has 512 dimensions, with each dimension being an integer from 0 to 255.

[0042] Step S2: Generate a 256-stream cipher key. Input the user-input password and salt value into a hash function, and output a pseudo-random number as the key. The length of the pseudo-random number is 512*(8+1)=4608 bits.

[0043] Step S3: Encrypt the original facial feature vector using a 256-stream cipher. Each dimension will eventually generate a pair of ciphertext, in the form of {<c1,s1> ,<c2,s2> ...<c512,s512> The encryption method is as follows: using the decimal feature vector generated from the facial information in step S1 and the key generated from the password in step S2, a 256-stream cipher is used for encryption. The input is the message m, and the keys key1 and key2 are used to mask one bit of p. The output ciphertext is a pair of...<c1,s1> The process of encrypting each dimension of the feature vector using a 256-stream cipher is as follows:

[0044] c1 = (m + key1) mod 256

[0045] p = (m + key1) > 255? 1:0

[0046]

[0047] Step S4: Face Similarity Calculation. Similarity calculation uses Euclidean distance. During the authentication phase, the Euclidean distance is calculated using the encrypted feature vectors from Step S3. Based on a threshold, it is determined whether they belong to the same person. If the Euclidean distance is less than or equal to the threshold, authentication is successful; otherwise, authentication fails. The Euclidean distance is obtained within the encrypted data. Specifically, the encrypted biometric feature vector has 512 dimensions, each an integer from 0 to 255. The Euclidean distance between two feature vectors is calculated as follows, with inputs (c1, s1)(c2, s2):

[0048] if Equal to 0: |E(c1)-E(c2)|

[0049] if Equals 1: 256 - |E(c1) - E(c2)|

[0050] The technological advancements and improved experimental data indicators of this embodiment are reflected in the following aspects: First, this embodiment protects biometric information without sacrificing recognition accuracy, achieving unchanged recognition performance before and after encryption. This is because the 256-stream cipher encryption scheme maintains the Euclidean distance. Second, this embodiment can run well on Android mobile devices, with a recognition speed higher than traditional homomorphic encryption, and it utilizes a cancelable biometric template and multiple fuzzy extractors.

[0051] Example 2

[0052] This embodiment provides an electronic device, including: one or more processors and a memory, wherein the memory stores one or more programs, the one or more programs including instructions for executing the stream cipher-based biometric authentication method as described in Embodiment 1.

[0053] Example 3

[0054] This embodiment provides a computer-readable storage medium including one or more programs executable by one or more processors of an electronic device, the one or more programs including instructions for performing the stream cipher-based biometric authentication method as described in Embodiment 1.

[0055] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A biometric authentication method based on stream ciphers, characterized in that, When applied to the server side, the following steps are included: Establish a connection with the client and obtain a preprocessed facial biometric vector from the client; Obtain the original password from the client and generate key information based on the original password; Based on the facial biometric vector and the key information, the encrypted text is obtained by encrypting it using a 256-stream cipher. For the encrypted ciphertext and the preset original biometric information, a similarity index is calculated between the two. It is then determined whether the similarity index is less than a preset threshold. If yes, authentication passes; otherwise, authentication fails and a prompt message is sent. The encrypted ciphertext is obtained using the following formula: ， ， ， In the formula, The value of a certain dimension in the facial biometric feature vector. and The values ​​that constitute the encrypted ciphertext in the corresponding dimension, The first key generated, Indicates modulo, when hour, ,when hour, , Used to cover a bit .

2. The biometric authentication method based on stream cipher according to claim 1, characterized in that, The process of obtaining the facial biometric vector includes the following steps: The facial biometric vector is obtained by sequentially processing the original image, performing face recognition and feature extraction via a deep convolutional neural network.

3. The biometric authentication method based on stream cipher according to claim 1, characterized in that, The process of obtaining the key information includes the following steps: For the original password, the obfuscated password information is obtained by adding salt, and a pseudo-random number is extracted using a preset random extractor as the key information.

4. The biometric authentication method based on stream cipher according to claim 1, characterized in that, The facial biometric vector is a decimal multidimensional vector.

5. The biometric authentication method based on stream cipher according to claim 1, characterized in that, The similarity metric mentioned is Euclidean distance.

6. The biometric authentication method based on stream cipher according to claim 1, characterized in that, For the form of ( , The encrypted ciphertext and its form are () , The acquisition of the similarity index, which is the original biological information of the organism, includes the following steps: judge ⊕ Is it equal to 0? If so, then the similarity index is... If not, then the similarity index is .

7. The biometric authentication method based on stream cipher according to claim 1, characterized in that, The facial biometric vector takes the value of an integer between 0 and 255 in each dimension.

8. An electronic device, characterized in that, include: One or more processors and a memory, wherein the memory stores one or more programs, the one or more programs being executed by the one or more processors, the one or more programs including instructions for performing the stream cipher-based biometric authentication method as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, It includes one or more programs that are executed by one or more processors of an electronic device, the one or more programs including instructions for performing the stream cipher-based biometric authentication method as described in any one of claims 1-7.

Images