CN114091094B - Fingerprint authentication and key negotiation method supporting updating - Google Patents

Abstract

The invention provides a fingerprint authentication and key negotiation method supporting updating, which is applied between a client and a server and comprises the following steps: initializing a system globally; registration: the client constructs a blinded certificate according to the input fingerprint information, initiates a registration request to the server based on the blinded certificate, and the server stores the blinded certificate and returns a registration result; authentication and key negotiation phase: the client initiates authentication to the server based on the fingerprint, the server generates a secret key and returns authentication reply information, the client inputs the correct fingerprint for authentication to acquire the secret key in the authentication reply information, and a communication channel between the client and the server is established on the basis of the secret key. The invention does not need extra encryption equipment to store the secret key or memorize the secret key, adopts fingerprint biological information to carry out authentication, and updates the blind certificate information facing the fingerprint when the certificate is destroyed so as to prevent further leakage.

Description

A fingerprint authentication and key negotiation method supporting updates

Technical Field

The present invention relates to the field of authentication and key negotiation, and in particular to a fingerprint authentication and key negotiation method supporting update.

Background Art

Existing authentication and key negotiation methods are usually implemented using traditional public-private key pairs or pre-shared passwords. These methods require devices to store private keys or users to remember passwords, which brings inconvenience to individual users. As a convenient identity recognition technology, biometrics has unique advantages in the field of authentication and key negotiation. However, existing biometric-based authentication and key negotiation schemes require the storage of user biometrics. Since user biometrics are sensitive personal data, extremely high requirements are placed on the secure storage technology of biometrics.

Summary of the invention

In view of the problems existing in the prior art, a fingerprint authentication and key negotiation method that supports updates is provided, so that users can realize authentication and key negotiation functions without memorizing information or holding special equipment, and the privacy of user biometrics does not rely on secure storage technology.

The technical solution adopted by the present invention is as follows:

A fingerprint authentication and key negotiation method supporting updates, applied between a client and a server, comprising:

System global initialization;

Registration phase: The client constructs a blind certificate based on the input fingerprint information, and initiates a registration request to the server based on the blind certificate. The server saves the blind certificate and returns the registration result.

Authentication and key negotiation phase: The client initiates authentication to the server based on the fingerprint. The server generates a key and returns an authentication reply message. The client enters the correct fingerprint for authentication to obtain the key in the authentication reply message, and establishes a communication channel with the server based on the key.

Furthermore, the process of constructing the blinded certificate is:

Step 21: Slice the input fingerprint information using a fingerprint slicing algorithm to obtain a fingerprint template information set;

Step 22: Perform fingerprint information hiding processing through the fingerprint template information set to obtain a blinded certificate.

Furthermore, in step 22, the hiding processing method is: using the discrete logarithm problem to hide the character string of the fingerprint template on the exponent.

Furthermore, the global initialization process is: executed by the server, outputting public parameters and a master key according to the input security parameters, the public parameters are sent to the client, and the master key is stored in the server.

Furthermore, when the client initiates a registration request, it sends a registration request message including a blinded certificate and an identity identifier to the server; the server receives the registration request message and returns a registration result based on the identity identifier, and at the same time saves the client identity identifier, update factor, auxiliary information, and blinded certificate in a database.

Furthermore, the specific process of the authentication and key agreement phase also includes:

The client inputs public parameters, fingerprint image information and basic fingerprint template information, processes the fingerprint image information to generate fingerprint template information; uses fingerprint matching algorithm to process to obtain new fingerprint template information; selects random fingerprint information from the new fingerprint template information to calculate a blind certificate, combines the blind certificate with the client identity to generate authentication request information and sends it to the server; and saves the status information locally on the client;

The server calculates auxiliary information to assist the client in generating a shared key; selects a secret token∈{0,1} ^λ and divides it into n small pieces; the server generates a secret key based on the server identity, client identity, secret, and blinded certificate. Generate a key and encrypt the key into authentication data; the server outputs authentication reply information and the key.

After receiving the authentication reply information, the client combines the public parameters and state information; uses its own correct fingerprint for authentication. If the authentication is successful, the secret credential information is recovered from the authentication reply information and combined with the state information to calculate the public session key.

Furthermore, the server receives the registration request information and returns the registration result according to the identity identifier. The specific process is: determine whether the identity identifier in the received registration request information has been registered. If not, output 1, initialize the update factor, and send it to the client. At the same time, save the update factor, identity identifier, blinding certificate, and auxiliary information in the database; otherwise, output 0 and reject the registration request.

Furthermore, the specific process of processing the fingerprint image information to generate the fingerprint template information is as follows: extracting n detail points from the fingerprint image information, and processing them into pieces to generate the fingerprint template information.

Furthermore, the fingerprint matching algorithm is: Is it true? If it is true, the fingerprint Add to the basic fingerprint template information to generate new fingerprint template information; is the fingerprint in the fingerprint template information, _ri is the fingerprint information in the basic fingerprint template, and τ is the distance threshold.

Furthermore, it also includes blind certificate update: if the master key is leaked, the server randomly generates a new master key to replace the original master key, and updates the public parameters and update factors, auxiliary information, and blind certificate in the database; if only the blind certificate is leaked, the blind certificate is directly updated.

Compared with the prior art, the beneficial effects of adopting the above technical solution are as follows: the present invention does not require additional encryption equipment to store keys or remember keys, and the use of fingerprint biometric information for construction provides a convenient tool for user identity authentication. Considering that the certificate is constructed based on fingerprint features, if it is not properly handled, it may leak the user's personal privacy information. The present invention fragments the fingerprint and represents it with a string, and performs blind processing. Considering that attackers may obtain data from the authentication and key negotiation process by invading the server or eavesdropping on the channel, the certificate may be stolen and misused. The present invention designs a blind certificate update mechanism, which allows the fingerprint-oriented blind certificate information to be updated when the certificate is damaged to prevent further leakage. The method can update the certificate database for self-recovery when an attack is detected.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of the fingerprint authentication and key negotiation process proposed by the present invention.

FIG. 2 is a schematic diagram of fingerprint representation proposed by the present invention.

DETAILED DESCRIPTION

The present invention will be further described below in conjunction with the accompanying drawings.

In view of the availability and cross-device convenience of user authentication and key negotiation processes, and the problem that fingerprint features involve user personal privacy, a fingerprint authentication and key negotiation method that supports updates is proposed, so that users can achieve authentication and key negotiation functions without having to remember information or hold special equipment, and the privacy of user biometrics does not rely on secure storage technology. The specific solution is as follows:

As shown in Figure 1, the system is globally initialized;

Registration phase: The client constructs a blind certificate based on the input fingerprint information, and initiates a registration request to the server based on the blind certificate. The server saves the blind certificate and returns the registration result.

Authentication and key negotiation phase: The client initiates authentication to the server based on the fingerprint. The server generates a key and returns an authentication reply message. The client enters the correct fingerprint for authentication to obtain the key in the authentication reply message, and establishes a communication channel with the server based on the key.

Specifically, in the process of constructing a blind certificate for fingerprints, an initialization process is first performed to input a security parameter λ and generate a public parameter par, including global parameters required for fingerprint information segmentation and an algebraic structure required for certificate generation, including a group G with a prime order of q, a generator g of the group G, and a threshold value τ set when calculating the Hamming distance.

After the initialization is completed, n detail points {point _i } _n are extracted from the input fingerprint image information, and the fingerprint segmentation algorithm is executed on the n detail points {point _i } _n . The character string set of the output fingerprint information template is shown in Figure 2, and the segmentation processing of the fingerprint information is completed.

In this embodiment, as shown in FIG2 , in order to accurately represent the fingerprint information, it is first necessary to extract n minutiae points {point _i } _n in the XY coordinate space. Specifically, each minutiae point is initialized as a center point, and then the γ points with the closest straight-line distance are selected. Assuming γ = 4, where point _i is the center point, we use Indicates the γ points closest to the center point. Indicates point _i to The vector relationship between two points, where d _j represents the vector Length, Represents the angular relationship between mutual vectors, so the fingerprint information is represented as a string Since fingerprint segmentation is based on the relative positions of minutiae points, rotating the fingerprint image does not affect the representation of the fingerprint.

After obtaining the character string set {r _i } _n of the fingerprint information template, fingerprint information hiding processing is performed on it to obtain the corresponding blinded certificate {b _i } _n .

Specifically, in this embodiment, the discrete logarithm problem is used to hide the character string of the fingerprint template on the exponent, that is, for the fingerprint information Hide _ri in In which h:{0,1} ^* →Z _q is a hash function, that is, a blinded certificate _bi is generated; Z _q is a residual class ring modulo q.

In this embodiment, a specific method of fingerprint-based authentication and key negotiation is also proposed, including three parts: global initialization, user registration, authentication and key negotiation.

Specifically, the global initialization is performed by the server, which inputs the security parameter λ, outputs the public parameter pp required by the entire system, and the master key information mk=(sk,sek), selects _sk∈Zq for confidentiality, and calculates ^gsk for public disclosure, where sek is the selected symmetric encryption key.

The public parameter pp also includes the group G of prime order q, the generator g in the group, the server's IDs, and the hash function H = (h ₀ ,h ₁ ,...,h _m-1 ) required in the obfuscated Bloom filter. The public parameter pp is received by the client, and the master key information mk is secretly stored by the server.

The user registration package has two parts: create registration and record registration

Create a registration:

This part is executed on the client side. By inputting the public parameter pp and the fingerprint image information finp sent by the server, the fingerprint image information finp is subjected to the fingerprint-oriented blind certificate construction process to obtain the corresponding blind certificate { _bi } _n , and then the registration request information Reg = {IDc, { _bi } _n } is generated. The client side sends the registration request information Reg to the server for registration, and at the same time, the fingerprint template information { _ri } _n generated in the process of constructing the blind certificate is secretly saved. It should be noted that in this embodiment, the client side must register with the server using the identity IDc and the fingerprint-based blind certificate { _bi } _n .

Record Registration:

This part is executed in the server. When the server receives the registration request information, it determines whether the identity IDc contained in the registration information Reg has been registered (i.e., whether it already exists in the database). If it has not been registered, it outputs 1, initializes an update factor Updc, and sends it to the client after being encrypted in a symmetric encryption form. At the same time, the server records (IDc, Updc, ω _c , {b _i } _n ) in the database DB to complete the registration; if it has been registered, it outputs 0 to reject the registration request. Among them, ω _c is the auxiliary information when updating the credential library, usually ω _c = g ^updc .

Specifically, the update factor Updc is set to 1 when initialized, and Updc is multiplied by a random value for updating when the credential database is updated.

Preferably, the symmetric encryption is performed using AES.

Authentication and key negotiation include creating a request, creating a response, and obtaining credentials. Specifically,

Create a request:

Enter the public parameter pp and fingerprint image information on the client And basic fingerprint template information { _ri } _n , from fingerprint image information Extract n detail points And it is processed into pieces to generate the corresponding fingerprint template information

Call the fingerprint matching algorithm for matching. In order to achieve high-precision fingerprint matching of multiple string templates, we use the Hamming distance to measure, that is, to calculate the number of different positions between two strings. The specific process is as follows: For the fingerprint template information set {r _i } _n and If there is a fingerprint information r _i ∈ {r _i } _n , Then calculate the Hamming distance if Where τ is the set distance threshold value, then Add to the set { _ri } _n ′.

The above fingerprint matching algorithm is used to obtain new fingerprint template information { _i } _n ′, random fingerprint information r _c ∈ Z _q is selected from the new fingerprint template information, and the blinded certificate is calculated. And the client temporarily saves the state locally Finally, the client generates authentication request information Send to the server.

Create a reply:

The server will expose the parameters pp, the information stored in the database DB, and the authentication request information And the master key information mk as input. To generate a response, the server first calculates the auxiliary information (where β∈Z _q ) to assist the client in generating a shared key. Then, a secret token∈{0,1} ^λ is selected. The length of the secret token is "λ", which is determined when the system is globally initialized, and is divided into n small blocks using secret sharing technology. In order to hide these n small blocks, we use a Garbled Bloom Filter (GBF). The key point is to Perform random segmentation and fill it into the confusion Bloom filter GBF by hashing { _i } _n . Finally, the server sets the key Where r _c , r _s ∈ Z _q , r _s is randomly generated by the server during key negotiation, and generates authentication encryption data c = AuthEnc (K _IDc , IDs). In order to prove its identity information to the client, the server generates η = r _s + l·sk (modq), where sk is the master key component. Finally, the server outputs the authentication reply information and the key

In this embodiment, the secret is divided into n small blocks corresponding to the n detail points in the fingerprint image, and the division into n small blocks can tolerate errors using the secret sharing method, that is, if t blocks out of the n small blocks are obtained, the secret token can be recovered, and t is the threshold of the secret sharing scheme.

Get the credentials:

After the client receives the authentication reply message Res from the server, it enters the public parameter pp and the status information saved locally The client uses its correct fingerprint for authentication. If the authentication is successful, it can reply with the secret That is, when the client fingerprint authentication succeeds, the client obtains the Credential information. At the same time, according to the value of _c of the locally saved status information, we can calculate Finally, we can get the value of At this point, the client already has the session key A secure communication channel is established with the server based on the

Finally, considering that attackers may obtain data from the authentication and key negotiation process by invading the server or eavesdropping on the channel, the certificate may be stolen and misused. In this embodiment, a blind certificate update mechanism is also proposed, which allows the fingerprint-oriented blind certificate information to be updated when the certificate is compromised to prevent further leakage. This method can update the certificate database for self-recovery when an attack is detected.

Once the server is compromised, the blinded certificate {b _i } _n and the master key mk will be leaked, and the blinded certificate update operation will be performed. Specifically,

When the master key is leaked, the server randomly generates a new master key mk′ to exclude the damaged key, and updates the public parameters pp and the information in the database DB. Specifically, a new key pair (g ^sk ′, sk′) is created, where sk′∈Z _q , and a new symmetric encryption key sek′∈Z _q is generated. As a result, the server updates the information Updc, ω _c , {b _i } _n in the data DB to Updc′, ω′ _c , {b _i }′ _n , respectively, and generates a new public parameter pp′ and master key mk′＝(sk′, sek′). The server can decide to update the database regularly, or update the database due to database damage. In this process, only the server needs to execute, and the client does not need to perform additional operation steps.

The specific database update algorithm is as follows: select α∈Z _q to update Updc, and calculate Updc′=Updc*α(mod q) and ω′ _c =g ^Updc ′, and finally output updated data ({b′ _i } _n , Updc′, ω′ _c ).

If only the blinded certificate { _i } _n is leaked and the master key mk is safe, then continue to maintain the original master key information mk and just re-execute the certificate update operation to generate a new blinded certificate.

The present invention is not limited to the aforementioned specific embodiments. The present invention extends to any new features or any new combination disclosed in this specification, as well as any new method or process steps or any new combination disclosed. If non-substantial changes or improvements made by those skilled in the art without departing from the spirit of the present invention should fall within the scope of protection of the claims of the present invention.

All features disclosed in this specification, or steps in all methods or processes disclosed, except mutually exclusive features and/or steps, can be combined in any manner.

Any feature disclosed in this specification, unless otherwise stated, can be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

Claims (8)

1. A fingerprint authentication and key agreement method supporting updating, applied between a client and a server, comprising:

Initializing a system globally;

Registration: the client constructs a blinded certificate according to the input fingerprint information, initiates a registration request to the server based on the blinded certificate, and the server stores the blinded certificate and returns a registration result;

Authentication and key negotiation phase: the client initiates authentication to the server based on the fingerprint, the server generates a secret key and returns authentication reply information, the client inputs the correct fingerprint for authentication to acquire the secret key in the authentication reply information, and a communication channel between the client and the server is established on the basis of the secret key;

The process of constructing the blinded certificate comprises the following steps: the input fingerprint information is subjected to slicing processing through a fingerprint slicing algorithm, so that a fingerprint template information set is obtained; fingerprint information hiding processing is carried out through the fingerprint template information set, and a blinded certificate is obtained;

the specific process of the authentication and key negotiation stage further comprises:

The client inputs the public parameters, the fingerprint picture information and the basic fingerprint template information, and processes the fingerprint picture information to generate fingerprint template information; processing by adopting a fingerprint matching algorithm to obtain new fingerprint template information; selecting random fingerprint information from the new fingerprint template information, calculating to obtain a blinded certificate, combining the blinded certificate with the client identity to generate authentication request information, and sending the authentication request information to a server; storing the state information locally at the client;

The server calculates auxiliary information to assist the client to generate a shared key; selecting a secret token epsilon {0,1} ^λ and dividing the secret token epsilon into n small blocks; the server generates a key based on the server identity, the client identity, the secret and the blinded certificate, and encrypts the key into authentication data; the server outputs authentication reply information and a secret key;

The client receives the authentication reply information and combines the public parameters and the state information; and adopting the correct fingerprint to carry out authentication, if the authentication is successful, recovering the secret credential information from the authentication reply information, and calculating by combining the state information to obtain the public session key.

2. The fingerprint authentication and key agreement method supporting updating according to claim 1, wherein the hiding processing method is: the character string of the fingerprint template is hidden on the index.

3. The method for supporting updated fingerprint authentication and key agreement according to claim 1, wherein the system global initialization procedure is: the method is executed by the server, public parameters and a master key are output according to the input security parameters, the public parameters are sent to the client, and the master key is stored in the server.

4. A fingerprint authentication and key agreement method supporting updating according to claim 1 or 3, wherein when a client initiates a registration request, registration request information including a blinded certificate and an identity is sent to a server; the server receives the registration request information and returns a registration result according to the identity, and simultaneously stores the client identity, the update factor, the auxiliary information and the blinded certificate in a database.

5. The fingerprint authentication and key agreement method supporting updating according to claim 1, wherein the specific process of the server receiving the registration request information and returning the registration result according to the identity is: judging whether the identity mark in the received registration request information is registered or not, if not, outputting 1, initializing an update factor, sending the update factor to a client, and simultaneously storing the update factor, the identity mark, the blinded certificate and auxiliary information in a database; and otherwise, outputting 0 and rejecting the registration request.

6. The method for supporting updated fingerprint authentication and key agreement according to claim 5, wherein the specific process of processing the fingerprint picture information to generate fingerprint template information is as follows: and extracting n minutiae from the fingerprint picture information, and performing fragmentation processing on the minutiae to generate fingerprint template information.

7. The method for supporting updated fingerprint authentication and key agreement according to claim 6, wherein the fingerprint matching algorithm is: judgingWhether or not it is true, if so, fingerprint is obtainedAdding the new fingerprint template information into the basic fingerprint template information to generate new fingerprint template information; wherein the method comprises the steps ofThe fingerprint template information is the fingerprint, r _i is the fingerprint information in the basic fingerprint template, and τ is the distance threshold value.

8. The method of fingerprint authentication and key agreement supporting updates according to claim 7, further comprising blinding certificate updates: if the master key is leaked, the server randomly generates a new master key to replace the original master key, and updates the public parameters, the update factors, the auxiliary information and the blinded certificate in the database; if only the blinded certificate is leaked, the blinded certificate is directly updated.