Face recognition based authentication key secure communication method and system

Abstract

This disclosure provides a secure communication method and system for authentication keys based on facial recognition, relating to the field of authentication key exchange technology in applied cryptography. The method includes initializing the two parties involved in the communication and generating common parameters; the two parties are a sender and a receiver; acquiring the facial information recorded by both parties, converting it into facial embedding vectors, and storing it in a set; generating a public key based on the recorded facial information; when the Euclidean distance between the facial embedding vector of the facial information and an element in the set is less than a set threshold, sending the public key to the other party through an authentication channel for exchange; mutual authentication and negotiation of a session key between the two parties; the sender selecting a random message to generate a session key, then encapsulating the random message; the receiver performing a facial registration to generate a private key to decapsulate the random message and obtain the session key, thus completing the acquisition of the random message. This disclosure achieves secure communication of facial recognition technology in high-dimensional noisy environments.

Description

Technical Field

[0001] This disclosure relates to the field of authentication key exchange technology in applied cryptography, specifically to a secure communication method and system for authentication keys based on facial recognition. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.

[0003] With the rapid development of the internet and mobile communication technologies, secure communication has gradually become one of the core issues in cyberspace. Globally, various applications and services are constantly emerging, and the transmission and storage of digital information has become an important part of daily life. Most secure communication services follow a store-and-forward model, where the communicating parties exchange information through a service intermediary. Considering the inability to guarantee the security of the store-and-forward process and the reliability of the service intermediary, encryption is used to protect messages, maximizing the prevention of attackers, including service intermediaries, from obtaining message content. This communication method is called end-to-end encryption. Specifically, participants authenticate themselves and negotiate a session key by performing an Authenticated Key Exchange (AKE), which in turn ensures the security of message transmission.

[0004] However, existing end-to-end encryption frameworks primarily rely on traditional public-key cryptography. This process requires each participant to generate a public-private key pair and distribute the public key to other participants. Then, the communicating parties execute a synchronous or asynchronous AKE protocol based on their respective public and private keys. However, these AKE protocols are not suitable for secure messaging in practice. First, these AKE protocols actually verify key ownership, not the participants themselves. Since keys are typically stored on the terminal device, attackers can directly perform chosen-ciphertext attacks. Second, once an attacker gains control of the terminal, it can lead to massive key leaks. More critically, service providers cannot effectively detect key cloning in a short time when such protocols are executed. Finally, updating the public-private key pair in real time becomes challenging when participants lose or replace their terminal devices, as each new public key requires out-of-band verification before activation.

[0005] Therefore, as facial recognition becomes increasingly popular globally as a means of authentication, researchers have begun to focus on authentication methods based on this biometric feature. Facial recognition, as a ubiquitous biometric feature, has been widely applied in fields such as face recognition and security authentication due to its advantages of uniqueness, universality, and convenience.

[0006] However, the inventors discovered that due to the permanent and irrevocable nature of facial features, personal privacy issues are becoming increasingly prominent. Facial recognition-based authentication key exchange, as an innovative authentication method, aims to achieve two-way authentication and key exchange without disclosing facial information. Despite its potential advantages, this method still faces challenges in practical applications, such as protecting facial information privacy, efficiency, and security. Summary of the Invention

[0007] To address the aforementioned issues, this disclosure proposes a secure communication method for authentication keys based on facial recognition, expanding the application scope of the BAKE framework and enabling it to complete authentication key exchange based on facial recognition. This allows facial recognition technology to be effectively used for secure communication in high-dimensional noise environments, providing a bidirectional biometric authentication key exchange (F-BAKE) scheme based on facial recognition technology for end-to-end secure communication.

[0008] According to some embodiments, the present disclosure adopts the following technical solutions:

[0009] A facial recognition-based authentication key secure communication method includes:

[0010] Initialize the communication participants and generate common parameters; the communication participants are the sender and the receiver.

[0011] Obtain the facial information entered by both parties involved in the communication, convert it into facial embedding vectors, and store it in a set;

[0012] A public key is generated based on the recorded facial information. When the Euclidean distance between the facial embedding vector of the facial information and an element in the set is less than a set threshold, the public key is sent to another participant through the authentication channel for exchange. The two participants authenticate each other and negotiate the session key.

[0013] Specifically, when the sender sends a random message, it selects the random message to generate a session key, then encapsulates the random message. The receiver performs a face scan to generate a private key to decapsulate the random message and obtain the session key, thus completing the acquisition of the random message.

[0014] According to some embodiments, the present disclosure adopts the following technical solutions:

[0015] A facial recognition-based authentication key secure communication system includes:

[0016] The initialization module is used to initialize the two parties involved in the communication and generate common parameters; the two parties involved in the communication are the sender and the receiver.

[0017] The registration module is used to obtain the facial information entered by both parties involved in the communication, convert it into facial embedding vectors, and store it in a set.

[0018] The key generation module is used to generate a public key based on the recorded facial information. When the Euclidean distance between the facial embedding vector of the facial information and an element in the set is less than a set threshold, the public key is sent to another participant through the authentication channel for exchange. The two participants authenticate each other and negotiate the session key.

[0019] The authentication key exchange module enables both participants to authenticate each other and negotiate session keys.

[0020] Specifically, when the sender sends a random message, it selects the random message to generate a session key, then encapsulates the random message. The receiver performs a face scan to generate a private key to decapsulate the random message and obtain the session key, thus completing the acquisition of the random message.

[0021] According to some embodiments, the present disclosure adopts the following technical solutions:

[0022] A non-transitory computer-readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the facial recognition-based authentication key secure communication method.

[0023] According to some embodiments, the present disclosure adopts the following technical solutions:

[0024] An electronic device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to perform the facial recognition-based authentication key secure communication method.

[0025] Compared with the prior art, the beneficial effects of this disclosure are as follows:

[0026] This disclosure provides a two-way biometric authentication key exchange (F-BAKE) framework based on facial recognition technology, offering a secure end-to-end communication solution that achieves secure and efficient end-to-end communication while fully leveraging existing facial recognition technology to improve the reliability and security of biometric authentication key exchange. The framework includes an initialization phase, a registration phase, a key generation phase, and an authentication key exchange phase. The initialization phase allows both participants to agree on some basic parameters and generate common parameters. The registration phase involves recording facial information and converting it into a facial embedding template, storing it in a set. Multiple recordings of facial information under different noise backgrounds (e.g., different angles, different lighting) are required to improve accuracy and robustness. The key generation phase generates a public key based on the recorded facial information and sends this public key to the other participant through an authentication channel when the Euclidean distance between the facial embedding vector and an element in the set is less than a threshold. The authentication key exchange phase enables both participants to authenticate each other and negotiate a session key. Considering that communication participants are typically not online simultaneously, and a participant may wish to send information to an offline participant through a service provider, this disclosure provides an asynchronous F-BAKE framework to address this asynchronous scenario. Subsequent session keys can be negotiated in the same manner, providing key rotation for the asynchronous F-BAKE framework. Key rotation is provided by independent session keys for each session. Even if an attacker successfully obtains a session key at some point, they cannot use that key to decrypt communication content in other sessions. This key rotation helps improve overall communication security. Attached Figure Description

[0027] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure.

[0028] Figure 1 The construction of decoding functions on D8 and E8 in embodiments of this disclosure.

[0029] Figure 2 This is a system model of the F-BAKE framework according to an embodiment of this disclosure.

[0030] Figure 3 This is a synchronous F-BAKE framework for embodiments of this disclosure.

[0031] Figure 4 This refers to the AKE phase of the asynchronous F-BAKE framework in this embodiment of the disclosure. Detailed Implementation

[0032] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.

[0033] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains.

[0034] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this disclosure. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0035] Example 1

[0036] One embodiment of this disclosure provides a secure communication method for authentication keys based on facial recognition, including:

[0037] Step 1: Initialize the two parties involved in the communication and generate common parameters; the two parties involved in the communication are the sender and the receiver;

[0038] Step 2: Obtain the facial information entered by both parties involved in the communication, convert it into facial embedding vectors, and store it in a set;

[0039] Step 3: Generate a public key based on the entered facial information. When the Euclidean distance between the facial embedding vector of the facial information and an element in the set is less than a set threshold, the public key is sent to another participant through the authentication channel for exchange.

[0040] Step 4: Both parties authenticate each other and negotiate the session key;

[0041] Specifically, when the sender sends a random message, it selects the random message to generate a session key, then encapsulates the random message. The receiver performs a face scan to generate a private key to decapsulate the random message and obtain the session key, thus completing the acquisition of the random message.

[0042] As one embodiment, the facial recognition-based authentication key secure communication method disclosed herein includes a two-way biometric authentication key exchange (F-BAKE) framework based on facial recognition technology. The framework consists of an E8-based security sketch, an E8-based fuzz extractor, an AFEM algorithm based on facial embedding vectors, a framework parameter initialization stage, a framework registration stage, a framework key generation stage, and a framework authentication key exchange stage.

[0043] This framework is based on F-AFEM and the standard BAKE framework.

[0044] F-AFEM includes four algorithms: Setup, PubGen, Enc, and Dec. The core of the PubGen algorithm is the use of an E8-based fuzz extractor, which can be derived from a secure sketch scheme based on E8 lattices using a stochastic language model, or a model based on LWE or DDH assumptions.

[0045] Among them, the E8-based security sketch protects sensitive information such as biometric data by providing error correction capabilities without revealing the original data. Used in the framework key generation stage, its basic principle is to generate supplementary information (Sketch) that can correct errors without disclosing the original data. This supplementary information can be used to recover the original data in noisy or corrupted input data environments, or to verify whether the input data matches the original data.

[0046] An E8-based fuzz extractor is used to extract stable and reliable keys from noisy raw data (such as biometrics). Because biometric data has inherent variability, biometric features collected at different times or under different environments may vary. The goal is to extract keys from this noisy data that can be used for encryption and authentication scenarios while protecting user privacy and data security.

[0047] A fuzz extractor is typically an extension of a basic random extractor, adding the ability to extract fuzzy data from noisy sources. Therefore, provided the input data has sufficiently high entropy, the output key of the fuzz extractor will be uniformly and randomly generated. Similarly, for noisy input data, if two input vectors differ only slightly under a certain metric, then the same key must be guaranteed for the output; if the two input vectors differ significantly, then the generated key should be completely unrelated to the reconstructed key.

[0048] The AFEM algorithm is based on facial embedding vectors. This algorithm is built on an E8 grid, meaning that the participant's facial information needs to be decoded into a real-valued vector, and after decomposition, it can be effectively decoded onto an E8 grid.

[0049] The framework initialization phase is used to enable both participants to reach an agreement on some basic parameters and generate common parameters that are accessible to both participants.

[0050] The frame registration phase is used to input facial information and convert it into facial embedding templates, which are then stored in a collection. This collection can be encrypted or stored directly in trusted memory. Multiple inputs of facial information under different noise backgrounds (e.g., different angles, different lighting) are required to improve accuracy and robustness.

[0051] The framework key generation phase is used to generate a public key based on the entered facial information. A public key is obtained when the Euclidean distance between the facial information embedding vector and an element in the set is less than a threshold; otherwise, facial information is re-entered until the test is passed. This public key is then sent to another participant through the authentication channel.

[0052] Furthermore, this embodiment is based on the E8 lattice, an 8-dimensional integer vector space lattice derived from the D8 lattice. It is the only positive definite, even, monomorphic lattice with a rank of 8. The E8 lattice consists of the D8 lattice and the set formed by adding each point to the vector (0.5, 0.5, ..., 0.5). In other words, the E8 lattice contains both all the points of the D8 lattice and the points obtained by adding the vector (0.5, 0.5, ..., 0.5) to each point of the D8 lattice. More figuratively, the E8 lattice can be viewed as a more compact structure formed by inserting an 8-dimensional semi-unit vector into each gap of the D8 lattice. A formalized definition of the E8 lattice is as follows:

[0053]

[0054] Here, the D8 lattice is an integer lattice in an 8-dimensional vector space, whose elements consist of integer vectors (x1, x2, ..., x8) with the following characteristics: each component x i Both belong to the set of integers Furthermore, the sum of the vector's components is even. The formal definition of the D8 lattice is as follows:

[0055]

[0056] Based on the relationship between E8 and D8 grids, a decoding function for the E8 grid can be constructed from the decoding function of the D8 grid. For the detailed algorithm flow, please see [link / description]. Figure 1 .

[0057] This embodiment, based on an E8-grid security sketch, generates additional information to correct errors without disclosing the original data. A security sketch (SS) with parameters (M, d, w, w′, m, m′, t) typically includes two probabilistic multinomial-time algorithms, where M is a metric space, d is a distance function under M, w and w′ are two biometric templates input in this metric space, m and m′ are the minimum entropy values ​​under different conditions, and t is a threshold parameter. The formal definition of the algorithm is as follows:

[0058] SS.Gen(w)→s: This function generates a security sketch s based on the original biometric template w. This sketch contains additional information to correct for potential errors and noise at the receiving end.

[0059] SS. This function performs error correction based on the new data sample w′ and the safety sketch s, and outputs the corrected feature template.

[0060] This embodiment is based on an E8 lattice fuzz extractor. A fuzz extractor with parameters (M, d, w, w′, m, t, λ) consists of three probabilistic multinomial-time algorithms: an initialization function (Init), a generation algorithm (Gen), and a reconstruction algorithm (Rep). The generation algorithm generates the key and auxiliary data based on the original data, while the reconstruction algorithm reconstructs the key using the auxiliary data and newly acquired noisy data. Here, M is a metric space, d is the distance function under M, w and w′ are two biometric templates input in this metric space, m is a minimum entropy value, t is a threshold parameter, and λ is the input security parameter. The formal definition of the fuzz extractor is as follows:

[0061] FE. The initialization function generates a common parameter pp based on the input security parameter λ, which is an implicit input to FE.Gen and FE.Rep.

[0062] FE. The generating function extracts a random string R and common information P from the input biometric features w∈M in order to recover R from another sample w′.

[0063] FE. The reconstruction function outputs the string R′ based on the input w′∈M and the common information P.

[0064] This embodiment is based on the AFEM algorithm using facial embedding vectors. Here, it is assumed that the key is... Where q is a prime number, F q It is a finite field. The specific algorithm definition and construction are as follows:

[0065] The PubGen algorithm's core public key generation process utilizes the aforementioned E8-based fuzz extractor to convert a noisy biometric vector into a public key for an ElGamal-like encryption algorithm. By running the FE.Gen algorithm, the biometric vector u is converted into a hashed extraction string R and a public string P. R is then hashed using the function H: Mapped to This generates a public key similar to the ElGamal algorithm. The public key at this point is (P, y = g). H(R) ), where g is a generator of the cyclic group.

[0066] The Enc algorithm: To encapsulate the message s, it utilizes an encryption process similar to the ElGamal algorithm. Specifically, for the public key y, a random value is chosen. The encapsulated message can be obtained.

[0067] The Dec algorithm: To decapsulate the received ciphertext message, it is also necessary to recover the key used in the ElGamal algorithm. Furthermore, the correctness of the fuzzy extractor and the construction of the E8-based fuzzy extractor guarantee that for all u, If the distance in a given metric space is less than the threshold parameter τ, then as long as FE.Gen(u)→(R,P), we can obtain FE.Rep(u′,P)→R.

[0068] Let sk, sk′∈F m For s∈G, a formalized definition of an F-AFEM is as follows:

[0069] Setup(1 λ Output par = (λ, G, q, g, τ, ι, H), where G is a prime-order cyclic group, g is a generator of G, τ is an integer, and H: It is a hash function.

[0070] PubGen(sk): Parses sk as u, runs FE.Gen(u)→(R,P) and calculates y=g. H(R) The output is pk = (P, y).

[0071] Enc(pk,s): Parses pk as (P,y). Selects a random value. Calculate c0 = g r and Output c = (P, c0, c1).

[0072] Dec(sk′,c): Resolves sk′ to u′ and c to (P,c0,c1). Run FE.Rep(u′,P)→R′. If y≠g H(R′) If the result is positive, output ⊥; otherwise, output [missing value].

[0073] like Figure 3 In this embodiment, the framework parameter initialization phase specifically includes the following steps:

[0074] Step A1: Participants P0 and P1 need to agree on some basic parameters, such as the safety parameter λ and the threshold τ.

[0075] Step A2: Call the AFEM.Setup algorithm to generate the common parameter par of AFEM.

[0076] Step A3: Set the public parameter of BAKE to par, making it accessible to both participants. In practical applications (such as secure communication), the service provider can generate the public parameter pp, publish it on a bulletin board or incorporate it into the software for participants to access.

[0077] like Figure 3 In this embodiment, during the framework registration phase, the participant's terminal needs to enter n (n∈N) * A set of facial information is generated and converted into a facial embedding template, which is then stored in a collection called SKA. This collection can be encrypted or stored directly in trusted memory. Multiple recordings of facial information are necessary because facial feature vectors, compared to Hamming biometrics such as fingerprints and irises, are typically located in more complex, high-dimensional noisy environments. Therefore, multiple recordings of facial information under different noise backgrounds (e.g., different angles, different lighting) are required to improve accuracy and robustness. The value of n depends on the embedding algorithm and hardware conditions. Ideally, n has a sigmoid-like relationship with the accuracy α.

[0078] like Figure 3 As shown, in the framework key generation stage of this embodiment, each participant generates a public key based on the entered facial information. The specific process includes the following steps:

[0079] Step B1: Participant P i (i∈{0,1}) Generate embedding vector sk i .

[0080] Step B2: Calculate the Euclidean distance to the SKA set elements. If any result is less than the threshold t, then run the AFEM.PubGen algorithm to obtain the public key pki; otherwise, re-enter the face information until the test is passed.

[0081] Step B3: P i Send the public key pki to P through the authentication channel. 1-i .

[0082] It is important to note that authentication channels are crucial for all authentication key exchange protocols, ensuring that public keys are not tampered with or forged. Currently, mainstream authentication channels can be implemented through public key infrastructures (PKIs), specifically by authoritative organizations generating certificates that bind identities to public keys. Furthermore, participants can also authenticate public keys by comparing public key fingerprints, scanning QR codes, and other methods.

[0083] In this embodiment, the framework authentication key exchange phase includes the following steps:

[0084] like Figure 3 As shown, in the synchronous F-BAKE framework:

[0085] Step C1: Participant P i (i∈{0,1}) First, select a random message s.i .

[0086] Step C2: Run the encapsulation algorithm AFEM.Enc to encapsulate s i Encapsulated as C i And send to P 1-i .

[0087] Step C3: Receive c 1-i Afterwards, P i Re-enter facial information to generate a private key sk′ i And use the decapsulation algorithm AFEM.Dec to decapsulate c 1-i Decapsulate to obtain s 1-i Then, P i Calculate the session key k using hash function H. i .

[0088] like Figure 4 As shown, in the asynchronous F-BAKE framework, assuming the first session is P0 online and P1 offline, its ID is sid0:

[0089] Step D1: Sender P0 first selects a random message. To generate session keys Then Packaged as

[0090] Step D2: When receiver P1 comes online, it will perform a face information registration and generate a private key for decryption. And obtain the session key

[0091] Step D3: Sender P1 initiates a second session sid1 and sends a message to receiver P0 containing the new session key. By decapsulation generate.

[0092] Subsequent session keys can be negotiated in the same way, providing key rotation for the asynchronous F-BAKE framework. This key rotation is provided by independent session keys for each session. Even if an attacker successfully obtains a session key at some point, they cannot use that key to decrypt communications in other sessions. This key rotation helps improve overall communication security.

[0093] Example 2

[0094] One embodiment of this disclosure provides a facial recognition-based authentication key secure communication system, including:

[0095] The initialization module is used to initialize the two parties involved in the communication and generate common parameters; the two parties involved in the communication are the sender and the receiver.

[0096] The registration module is used to obtain the facial information entered by both parties involved in the communication, convert it into facial embedding vectors, and store it in a set.

[0097] The key generation module is used to generate a public key based on the recorded facial information. When the Euclidean distance between the facial embedding vector of the facial information and an element in the set is less than a set threshold, the public key is sent to another participant through the authentication channel for exchange. The two participants authenticate each other and negotiate the session key.

[0098] The authentication key exchange module enables both participants to authenticate each other and negotiate a session key;

[0099] Specifically, when the sender sends a random message, it selects the random message to generate a session key, then encapsulates the random message. The receiver performs a face scan to generate a private key to decapsulate the random message and obtain the session key, thus completing the acquisition of the random message.

[0100] Example 3

[0101] One embodiment of this disclosure provides a non-transitory computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the facial recognition-based authentication key secure communication method.

[0102] Example 4

[0103] One embodiment of this disclosure provides an electronic device, including: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to perform the facial recognition-based authentication key secure communication method.

[0104] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0105] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0106] While the specific embodiments of this disclosure have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of this disclosure. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of this disclosure are still within the scope of protection of this disclosure.

Claims

1. A secure communication method for authentication keys based on facial recognition, characterized in that, include: Initialize the communication between the two parties and generate common parameters; The two parties involved in the communication are the sender and the receiver; Obtain the facial information entered by both parties involved in the communication, convert it into facial embedding vectors, and store it in a set; A public key is generated based on the recorded facial information. When the Euclidean distance between the facial embedding vector of the facial information and an element in the set is less than a set threshold, the public key is sent to another participant through the authentication channel for exchange. The two participants authenticate each other and negotiate the session key. When the Euclidean distance between the face embedding vector of the face information and an element in the set is less than a set threshold, the public key is sent to another participant through the authentication channel for exchange, including: exchanging authentication keys in a synchronous F-BAKE or asynchronous F-BAKE framework; The F-BAKE framework is based on a combination of F-AFEM and the standard BAKE framework; F-AFEM includes four algorithms: Setup, PubGen, Enc, and Dec; The F-BAKE framework consists of an E8-based security sketch, a fuzz extractor, an AFEM algorithm based on face embedding vectors, a framework parameter initialization phase, a framework registration phase, a framework key generation phase, and a framework authentication key exchange phase. Specifically, when the sender sends a random message, it selects the random message to generate a session key, then encapsulates the random message. The receiver performs a face scan to generate a private key to decapsulate the random message and obtain the session key, thus completing the acquisition of the random message.

2. The facial recognition-based authentication key secure communication method as described in claim 1, characterized in that, Initialize the communication between the two parties and generate common parameters, including: The two parties reach an agreement based on the basic parameters, call the AFEM.Setup algorithm to generate the common parameters of AFEM, and set the common parameters of BAKE for access by both participants; The basic parameters are safety parameters and set thresholds.

3. The facial recognition-based authentication key secure communication method as described in claim 1, characterized in that, Participant terminals need to input multiple facial information entries and convert them into facial embedding vectors, which are then stored in a set. The system records facial information from multiple people under different noise backgrounds; the collection can be encrypted and saved or stored directly in trusted memory.

4. The facial recognition-based authentication key secure communication method as described in claim 3, characterized in that, A public key is generated based on the facial information entered by the participants, including: Step 1: Participants Generate embedding vectors ; Step 2: Calculation and The Euclidean distance between the elements of the set; if any result is less than the threshold. If so, run the AFEM.PubGen algorithm to obtain the public key; otherwise, re-enter the face information until the test is passed. Step 3: Send the public key to the authentication channel. .

5. The facial recognition-based authentication key secure communication method as described in claim 1, characterized in that, The specific process of authentication key exchange includes the following steps: Step C1: Participants First, select a random message. ; Step C2: Run the encapsulation algorithm AFEM.Enc to... Packaged as And send to ; Step C3: Received back, Re-enter facial information to generate a private key And use the decapsulation algorithm AFEM.Dec to... Decapsulate to obtain ;Then, Through hash function Calculate session key .

6. The facial recognition-based authentication key secure communication method as described in claim 1, characterized in that, In the asynchronous F-BAKE framework, assuming the first session is Online When offline, its ID is : Step D1: Sender First, select a random message. To generate session keys Then Packaged as ; Step D2: When the receiver Upon going live, he will perform a facial recognition scan and generate a private key for decryption. And obtain the session key ; Step D3: Sender Start a second session to the recipient Send a message containing the new session key. By decapsulation generate.

7. A facial recognition-based authentication key secure communication system, characterized in that, include: The initialization module is used to initialize the two parties involved in the communication and generate common parameters; The two parties involved in the communication are the sender and the receiver; The registration module is used to obtain the facial information entered by both parties involved in the communication, convert it into facial embedding vectors, and store it in a set. The key generation module is used to generate a public key based on the recorded facial information. When the Euclidean distance between the facial embedding vector of the facial information and an element in the set is less than a set threshold, the public key is sent to another participant through the authentication channel for exchange. The two participants authenticate each other and negotiate the session key. When the Euclidean distance between the face embedding vector of the face information and an element in the set is less than a set threshold, the public key is sent to another participant through the authentication channel for exchange, including: exchanging authentication keys in a synchronous F-BAKE or asynchronous F-BAKE framework; The F-BAKE framework is based on a combination of F-AFEM and the standard BAKE framework; F-AFEM includes four algorithms: Setup, PubGen, Enc, and Dec; The F-BAKE framework consists of an E8-based security sketch, a fuzz extractor, an AFEM algorithm based on face embedding vectors, a framework parameter initialization phase, a framework registration phase, a framework key generation phase, and a framework authentication key exchange phase. The authentication key exchange module enables both participants to authenticate each other and negotiate a session key; Specifically, when the sender sends a random message, it selects the random message to generate a session key, then encapsulates the random message. The receiver performs a face scan to generate a private key to decapsulate the random message and obtain the session key, thus completing the acquisition of the random message.

8. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the facial recognition-based authentication key secure communication method as described in any one of claims 1-6.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to perform the secure communication method based on facial recognition authentication keys as described in any one of claims 1-6.

Images