A multi-modal identity authentication system based on multiple pulse waves and veins and an authentication method thereof

Abstract

The application discloses a multi-modal identity authentication system based on multiple pulse waves and veins and an authentication method thereof, comprising a data acquisition module, a data processing module and a multi-modal authentication module. The multi-modal identity authentication method stores user dorsum manus vein information and contact finger pulse wave information to a server in a registration stage, collects the dorsum manus vein information and the contact finger pulse wave information again in an authentication stage, and checks and authenticates the dorsum manus vein information and the contact finger pulse wave information stored in the server. Meanwhile, non-contact dorsum manus pulse wave and contact finger pulse wave are subjected to double pulse wave similarity verification. Through multiple identity recognitions of the dorsum manus vein feature, the contact finger pulse wave feature and the double pulse wave similarity verification, the security of identity recognition is improved, and the function of living body detection is simultaneously achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of identity authentication, and in particular to a multimodal identity authentication system based on multiple pulse waves and veins and an authentication method thereof. Background Art

[0002] With the rapid development of the economy and society, the importance of information security has become increasingly apparent, and the harm caused by identity theft has become more serious. As a key means of protecting information security, identity authentication technology has increasingly high requirements for its universality, specificity, and stability. Traditional identity authentication methods such as passwords and ID cards are no longer able to meet people's needs due to their shortcomings such as being easily copied, stolen, or forgotten. Biometric technology has emerged as a response to this need.

[0003] Compared to traditional identity authentication technologies, biometrics offers significant advantages, including impermeability, portability, and high security. Vein recognition, which uses the pattern of veins in the human body to identify individuals, holds significant potential. However, similar to facial recognition, attacks against it are constantly emerging, such as the use of photos and videos to easily deceive vein recognition systems.

[0004] Embedding liveness detection features in facial recognition, such as prompts for blinking and opening the mouth, can effectively prevent this type of attack. However, vein recognition is not easily embedded with motion detection for liveness detection. Furthermore, single-modality vein recognition suffers from the drawback of single-algorithm authentication.

[0005] Biometrics based on heart signals have many advantages over traditional biometrics:

[0006] 1) Intrinsic activity: Heart movement only exists in "live" users, and the system can distinguish whether the verification object is a "live" user or a forged attack;

[0007] 2) Highly safe: cardiac signals depend on the user's heart muscle structure and therefore cannot be completely mimicked;

[0008] Physiological signals from the human heart are economically efficient and offer the potential for feature recognition, holding great potential for biometrics. The pulse wave signal, a periodic fluctuation in light absorption caused by changes in arterial blood flow with heartbeat, is easily acquired, making it suitable as a physiological signature for biometric systems.

[0009] However, compared with more mature identity authentication technologies, the performance of single pulse wave recognition algorithms is poor. Summary of the Invention

[0010] The present invention aims to address the deficiencies of the prior art and to propose a multimodal identity authentication system and an authentication method based on multiple pulse waves and veins.

[0011] The object of the present invention is achieved through the following technical solutions: a multimodal identity authentication system based on multiple pulse waves and veins, the system comprising: a data acquisition module, a data processing module and a multimodal authentication module;

[0012] The data acquisition module includes a contact pulse wave acquisition module and a hand back video acquisition module; the contact pulse wave acquisition module is used to acquire contact finger pulse waves; the hand back video acquisition module is used to acquire non-contact hand back pulse waves and hand back vein images;

[0013] The data processing module includes a contact finger pulse wave feature extraction module, a back hand vein feature extraction module and a non-contact back hand pulse wave acquisition module; the contact finger pulse wave feature extraction module is used to pre-process and extract features of the contact finger pulse wave signal acquired by the contact pulse wave acquisition module to obtain the user's contact finger pulse wave feature vector; the back hand vein feature extraction module is used to pre-process and extract features of the back hand image acquired by the back hand video acquisition module, first acquiring the back hand vein image, and then acquiring the user's back hand vein feature vector based on the vein image; the non-contact back hand pulse wave acquisition module is used to acquire the non-contact back hand pulse wave based on the back hand video information acquired by the back hand video acquisition module;

[0014] The multimodal authentication module includes a contact finger pulse wave feature authentication module, a back hand vein feature authentication module, and a dual pulse wave similarity verification module; the contact finger pulse wave feature authentication module is used to compare and verify the similarity between the feature vector of the user's second contact finger pulse wave signal and the feature vector of the first contact finger pulse wave signal; the back hand vein feature authentication module is used to compare and verify the similarity between the feature vector of the user's second back hand vein image and the feature vector of the first back hand vein image; the dual pulse wave similarity verification module is used to perform similarity analysis on the user's contact finger pulse wave signal and non-contact back hand pulse wave signal to determine whether the dual pulse waves belong to the same authentication subject; and a final identity authentication result is generated based on the joint verification results of the contact finger pulse wave feature authentication module, the back hand vein feature authentication module, and the dual pulse wave similarity verification module.

[0015] Furthermore, the contact pulse wave acquisition module includes an infrared pulse wave sensor, which includes an infrared finger clip probe and an acquisition module for acquiring contact finger pulse wave signals.

[0016] Furthermore, the back of the hand video acquisition module includes a near-infrared LED lamp and a near-infrared camera; the near-infrared LED lamp serves as the light source for the near-infrared camera when shooting; the near-infrared camera is used to record the user's back of the hand video and back of the hand image, and further obtain the non-contact pulse wave and back of the hand vein image of the back of the hand.

[0017] Furthermore, the non-contact hand back pulse wave acquisition module acquires the non-contact hand back pulse wave based on the near-infrared hand back video collected by the hand back video acquisition module, specifically as follows:

[0018] 1) Frame the near-infrared hand back video;

[0019] 2) Perform vein detection on each frame of the hand back image to obtain a vein image;

[0020] 3) averaging the pixels in the vein area of ​​each frame;

[0021] 4) constructing a set of time-varying sequences based on the pixel averages obtained in step 3) according to the video sequence;

[0022] 5) The sequence obtained in step 4) is filtered to obtain a pulse wave.

[0023] Furthermore, the pulse wave signal preprocessing and feature extraction are specifically as follows: performing wavelet threshold denoising on the collected pulse wave; performing peak detection on the denoised pulse wave signal to segment the single-cycle pulse wave; performing kernel principal component analysis and principal component analysis on the single-cycle pulse wave to complete dimensionality reduction of the pulse wave signal and obtain a feature vector as the first contact finger pulse wave sample.

[0024] Furthermore, the similarity comparison between the contact-type finger pulse wave feature authentication module and the hand back vein feature authentication module is specifically as follows: the pulse wave (or hand back vein) feature vector X (x1, x2, ..., x n ) and the pulse wave (or hand vein) feature vector Y1 (y 11 ,y 12 ,...,y 1n ), Y2…Y m The Euclidean distance of:

[0025]

[0026] The k-Nearest Neighbor (KNN) classifier is used for classification recognition and similarity comparison.

[0027] Furthermore, the similarity analysis of the contact finger pulse wave signal and the non-contact back of hand pulse wave signal is specifically as follows: the finger pulse wave signal is recorded as P1, and the back of hand pulse wave signal is recorded as P2; P1 and P2 are normalized in amplitude and time, recorded as P1′ and P2′; then the Euclidean distance d(P1′, P2′) between P1′ and P2′ is calculated. If d(P1′, P2′) is less than a threshold, it is determined that P1 and P2 belong to the same subject.

[0028] The present invention also provides a multimodal identity authentication method based on multiple pulse waves and veins. The method is divided into a registration phase and an authentication phase, and includes the following steps:

[0029] During the registration phase, S1 uses a near-infrared camera to capture the back of the registered user's hand image to obtain a first back of the hand vein image, and stores the first back of the hand vein image in the database as authentication reference information;

[0030] S2, during the registration phase, uses a contact finger pulse wave sensor to collect a contact finger pulse wave signal to obtain a first contact finger pulse wave sample, and stores the first contact finger pulse wave sample in a database as authentication reference information;

[0031] During the authentication phase, S3 uses a near-infrared camera to collect video information of the back of the hand of the authenticated user to obtain a second back-hand vein image and a non-contact back-hand pulse wave signal;

[0032] In step S4, during the authentication phase, the contact finger pulse wave sensor is used to collect the contact finger pulse wave signal of the authenticated user to obtain a second contact finger pulse wave sample. During the collection of the second contact finger pulse wave sample, the back of the hand video information is continuously obtained through the near-infrared camera. After the contact finger pulse wave signal is collected, step S5 is executed.

[0033] S5 performs vein recognition authentication by comparing the second hand back vein image obtained in step S3 with the first hand back vein image by similarity;

[0034] S6: performing contact finger pulse wave authentication by comparing the second contact finger pulse wave sample obtained in step S4 with the first contact finger pulse wave sample through similarity;

[0035] S7 performs dual pulse wave similarity verification on the contact finger pulse wave signal and the non-contact hand back pulse wave signal obtained in step S4;

[0036] S8 generates an identity authentication result based on the combined verification results of hand back vein recognition authentication, contact pulse wave authentication and dual pulse wave similarity verification.

[0037] Furthermore, the specific process of step S1 is: collecting the back of the hand image information of the registered user, detecting whether the collected image is the back of the hand through the back of the hand detection algorithm, and if so, saving it as the first back of the hand vein image, otherwise re-collecting the image.

[0038] The beneficial effects of the present invention are:

[0039] (1) The present invention combines hand vein recognition authentication, pulse wave authentication and dual pulse wave similarity verification to achieve multimodal identity authentication, which improves the security of the algorithm and solves the shortcomings of single algorithm authentication.

[0040] (2) The present invention utilizes contact pulse waves for identity authentication and non-contact pulse waves for dual pulse wave verification, and has the function of liveness detection.

[0041] (3) After the present invention completes the collection of biometric information (the second hand back vein image, the second contact finger pulse wave sample, and the non-contact hand back pulse wave signal), it directly completes multimodal identity authentication and liveness detection in combination with the pre-registered information (the first hand back vein image and the first contact finger pulse wave sample), and the identity authentication and liveness detection links are unified. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a registration process flowchart of a multimodal identity authentication method based on multiple pulse waves and hand back veins.

[0043] Figure 2 This is a flowchart of an authentication process based on a multimodal identity authentication method using multiple pulse waves and hand veins.

[0044] Figure 3 This is the overall structural block diagram of a multimodal identity authentication system based on multiple pulse waves and hand back veins.

[0045] Figure 4 Flowchart of the algorithm for contactless hand dorsum pulse wave acquisition.

[0046] Figure 5 Schematic diagram of a multimodal identity authentication system based on multiple pulse waves and hand back veins.

[0047] Figure 6 This is a schematic diagram of hand data acquisition for a multimodal identity authentication system based on multiple pulse waves and hand veins, where a) is a front view of the hand data acquisition device and b) is a top view of the hand data acquisition device. DETAILED DESCRIPTION

[0048] In order to illustrate in more detail a multimodal identity authentication system and authentication method based on multiple pulse waves and veins of the present invention, the present invention is described in detail below with reference to the accompanying drawings.

[0049] Example 1

[0050] like Figure 1 and Figure 2 As shown, this embodiment provides a multimodal identity authentication method based on multiple pulse waves and veins, which includes the following steps:

[0051] During the registration phase, S1 uses a near-infrared camera to capture the back of the registered user's hand image to obtain a first back of the hand vein image, and stores the first back of the hand vein image in the database as authentication reference information;

[0052] S2, during the registration phase, uses a contact finger pulse wave sensor to collect a contact finger pulse wave signal to obtain a first contact finger pulse wave sample, and stores the first contact finger pulse wave sample in a database as authentication reference information;

[0053] During the authentication phase, S3 uses a near-infrared camera to collect video information of the user's hand back to obtain a second hand back vein image and a non-contact hand back pulse wave signal. The method of extracting pulse wave data based on video can refer to the authorized patent "A non-contact pulse wave acquisition system and acquisition method for non-constrained targets" (201310082729.3).

[0054] During the authentication phase, step S4 uses the contact finger pulse wave sensor to collect the contact finger pulse wave signal of the authenticated user to obtain a second contact finger pulse wave sample. During the collection of the contact finger pulse wave sample, the back of the hand image is continuously acquired. If the back of the hand image cannot be acquired, the process returns to step S3. After the contact finger pulse wave signal acquisition is completed, step S5 is executed.

[0055] S5 performs vein recognition authentication on the second hand back vein image obtained in the above step;

[0056] S6: performing contact finger pulse wave authentication on the second contact finger pulse wave sample obtained in the above step;

[0057] S7 performs dual pulse wave similarity verification on the contact finger pulse wave signal and the non-contact hand back pulse wave signal obtained in the above step;

[0058] S8 generates an identity authentication result based on the results of hand vein recognition authentication, contact pulse wave authentication and dual pulse wave similarity verification.

[0059] The first contact finger pulse wave sample in step S2 is specifically: performing feature recognition on the collected contact finger pulse wave signal, weighting each feature to form a feature vector as the first contact finger pulse wave sample and storing it in the database.

[0060] The hand back vein recognition authentication described in step S5 is specifically as follows: all the second hand back vein images obtained in steps S3 to S4 are compared one by one with the first hand back vein images stored in the database in step S1. If they are consistent, the hand back recognition authentication is passed. If they are inconsistent, the authentication is failed.

[0061] The contact finger pulse wave authentication in step S6 is specifically as follows: all the second contact finger pulse wave samples obtained in step S4 are compared one by one with the first contact finger pulse wave samples stored in the database in step S2. If they are consistent, the pulse wave authentication is passed. If they are inconsistent, the authentication is failed.

[0062] The dual pulse wave similarity verification described in step S7 is specifically as follows: performing similarity analysis on the contact finger pulse wave signal obtained in step S4 and the non-contact hand back pulse wave signal obtained in step S3. If the similarity is higher than a threshold, the dual pulse wave similarity verification is passed.

[0063] Step S8 is specifically as follows: if the hand vein recognition authentication, pulse wave authentication and double pulse wave similarity verification are all passed, the identity authentication is determined to be passed; otherwise, it is determined to be failed.

[0064] Example 2

[0065] like Figure 3 As shown, this embodiment provides a multimodal identity authentication system based on multiple pulse waves and veins, the system comprising:

[0066] The data acquisition module includes a contact pulse wave acquisition module and a hand back video acquisition module.

[0067] The data processing module includes a contact finger pulse wave feature extraction module, a hand back vein feature extraction module and a non-contact hand back pulse wave acquisition module.

[0068] The multimodal authentication module includes a contact-type finger pulse wave feature authentication module, a hand back vein feature authentication module, and a dual pulse wave similarity verification module.

[0069] The contact pulse wave acquisition module is used to acquire contact finger pulse waves and includes an infrared pulse wave sensor, a standard USB interface, and a power supply. The infrared pulse wave sensor includes an infrared finger clip probe and an acquisition module for acquiring contact finger pulse wave signals. The standard USB interface is used to transmit pulse wave signals. The power supply is used to power the USB.

[0070] The hand back video acquisition module includes a near-infrared LED light, a near-infrared camera, and a standard USB port. The near-infrared LED light serves as a light source; the near-infrared camera records videos and images of the user's hand back, further acquiring non-contact pulse waves and hand veins; and the standard USB port transmits hand back videos. For methods of extracting pulse wave data from video, refer to the authorized patent "A Non-Contact Pulse Wave Acquisition System and Method for Unconstrained Targets" (201310082729.3).

[0071] The contact finger pulse wave feature extraction module is used to preprocess and extract features from the contact finger pulse wave signal acquired by the contact pulse wave acquisition module to acquire a user's contact finger pulse wave feature vector;

[0072] The hand back vein feature extraction module is used to preprocess and extract features from the hand back image acquired by the hand back video acquisition module, first acquiring the hand back vein image, and then acquiring the user's hand back vein feature vector based on the vein image;

[0073] The non-contact hand back pulse wave acquisition module acquires the non-contact hand back pulse wave based on the near infrared hand back video collected by the hand back video acquisition module. Figure 4 As shown, the specific algorithm is:

[0074] 1) Frame the near-infrared hand back video;

[0075] 2) Perform vein detection on each frame of the hand back image to obtain a vein image. The specific method of vein detection can be found in "A Near-Infrared Hand Back Vein Image Recognition Algorithm" (DOI: 10.19304 / j.cnki.issn1000-7180.2010.10.028);

[0076] 3) averaging the pixels in the vein area of ​​each frame;

[0077] 4) constructing a set of time-varying signals from the pixel averages obtained in step 3) according to the video sequence;

[0078] 5) Processing the signal obtained in step 4) to obtain a pulse wave. The specific processing method can refer to the authorized patent "A non-contact pulse wave acquisition system and acquisition method for unconstrained targets" (201310082729.3).

[0079] The contact finger pulse wave feature authentication module compares the similarity between the feature vector of the user's second contact finger pulse wave signal and the feature vector of the first contact finger pulse wave signal, uses a k-nearest neighbor classifier (KNN) to classify and identify the features, and selects Euclidean distance as the similarity metric for identity recognition, which is used as part of user identity verification;

[0080] The hand back vein feature authentication module refers to a part of the user identity verification that compares the similarity between the feature vector of the second hand back vein image of the user and the feature vector of the first hand back vein image;

[0081] The dual pulse wave similarity verification module refers to a part of user identity verification that performs similarity analysis on the user's contact finger pulse wave signal and non-contact hand back pulse wave signal to determine whether the dual pulse waves belong to the same authentication subject;

[0082] The pulse wave signal preprocessing specifically includes: performing wavelet threshold denoising on the collected pulse wave; performing peak detection on the denoised pulse wave signal, and segmenting the single-cycle pulse wave.

[0083] The pulse wave signal feature extraction specifically includes: performing kernel principal component analysis and principal component analysis on a single-cycle pulse wave to complete dimensionality reduction of the pulse wave signal and extract a feature vector with resolution.

[0084] The similarity comparison between the contact finger pulse wave feature authentication module and the hand back vein feature authentication module is specifically as follows: the pulse wave (or hand back vein) feature vector X (x1, x2, ..., x n ) and the pulse wave (or hand vein) feature vector Y1 (y 11 ,y 12 ,…,y 1n ), Y2…Y m The Euclidean distance of:

[0085]

[0086] The k-Nearest Neighbor (KNN) classifier is used for classification recognition and similarity comparison.

[0087] Furthermore, the similarity analysis of the contact finger pulse wave signal and the non-contact back of the hand pulse wave signal is specifically as follows: the finger pulse wave signal is recorded as P1, and the back of the hand pulse wave signal is recorded as P2; P1 and P2 are normalized in amplitude and time, recorded as P1′ and P2′; then the Euclidean distance d(P1′, P2′) between P1′ and P2′ is calculated. If d(P1′, P2′) is less than the threshold, it is determined that P1 and P2 belong to the same subject.

[0088] Furthermore, the threshold calculation process is specifically as follows: let the finger pulse wave signal of subject 1 be P 11 , the hand back pulse wave signal is P 12 ; The finger pulse wave signal of subject 2 is P 21 , the hand back pulse wave signal is P 22 ; The finger pulse wave signal of subject n is P n1 , the hand back pulse wave signal is P n2 ; for P 11 , P 12 , P 21 , P 22 ,...,P n1 , P n2 Normalize the amplitude and time, denoted as P 11 ′、P 12 ′、P 21 ′、P 22 ′,…,P n1 ′、P n2 '; then calculate the Euclidean distance d (P 11 ′,P 12 ′), d(P 21 ′,P 22 ′),…,d(P n1 ′,P n2 ′), the maximum value of which is denoted as d max ; Then calculate the Euclidean distance d (P 11 ′,P 12 ′), d(P 21 ′,P 22 ′),…,d(P n1 ′,P n2 ′), the minimum value of which is denoted as d min ; The threshold is

[0089] Example 3

[0090] like Figure 5 and Figure 6As shown, the following will be taken as an example of an attacker holding a user's hand back vein video for a forgery attack to specifically explain a multi-modal identity authentication system based on multiple pulse waves and veins and an authentication method thereof.

[0091] The attacker holds a user's hand back vein video for a forgery attack, and the main steps in the identity authentication stage are as follows:

[0092] 1) The hand data acquisition device (near-infrared camera) is used to collect the hand back video information of the authentication user, and the second hand back vein image and the non-contact hand back pulse wave signal are obtained.

[0093] 2) The contact type finger pulse wave sensor is used to collect the contact type finger pulse wave signal of the attacker, and the second contact type finger pulse wave sample is obtained.

[0094] 3) The second hand back vein image obtained in the above step is subjected to vein recognition authentication; since the attacker attacks by using the hand back video of the authentication user, the vein recognition authentication passes;

[0095] 4) The second contact type finger pulse wave sample obtained in the above step is subjected to contact type finger pulse wave authentication; since the second contact type finger pulse wave sample obtained in the attack process is the pulse wave of the attacker, the contact type finger pulse wave authentication does not pass;

[0096] 5) The contact type finger pulse wave signal and the non-contact hand back pulse wave signal obtained in the above step are subjected to double pulse wave similarity verification; since the contact type finger pulse wave signal obtained in the above step is the pulse wave of the attacker, and the non-contact hand back pulse wave signal is the pulse wave of the authentication user obtained by the hand back video of the authentication user, the double pulse wave similarity verification does not pass;

[0097] 6) According to the results of the hand back vein recognition authentication, the contact type pulse wave authentication and the double pulse wave similarity verification, the final identity authentication does not pass.

[0098] The system described in the application can successfully prevent photo forgery attacks in hand back vein recognition.

[0099] The above examples are used to explain and illustrate the application, but not to limit the application, and any modifications and changes made to the application within the spirit and protection scope of the claims fall within the protection scope of the application.

Claims

1. A multimodal identity authentication system based on multiple pulse waves and veins, characterized in that: The system includes: a data acquisition module, a data processing module and a multimodal authentication module; The data acquisition module includes a contact pulse wave acquisition module and a hand back video acquisition module; the contact pulse wave acquisition module is used to acquire contact finger pulse waves; the hand back video acquisition module is used to acquire non-contact hand back pulse waves and hand back vein images; The data processing module includes a contact finger pulse wave feature extraction module, a back hand vein feature extraction module and a non-contact back hand pulse wave acquisition module; the contact finger pulse wave feature extraction module is used to pre-process and extract features of the contact finger pulse wave signal acquired by the contact pulse wave acquisition module to obtain the user's contact finger pulse wave feature vector; the back hand vein feature extraction module is used to pre-process and extract features of the back hand image acquired by the back hand video acquisition module, first acquiring the back hand vein image, and then acquiring the user's back hand vein feature vector based on the vein image; the non-contact back hand pulse wave acquisition module is used to acquire the non-contact back hand pulse wave based on the back hand video information acquired by the back hand video acquisition module; The multimodal authentication module includes a contact finger pulse wave feature authentication module, a back hand vein feature authentication module, and a dual pulse wave similarity verification module; the contact finger pulse wave feature authentication module is used to compare and verify the similarity between the feature vector of the user's second contact finger pulse wave signal and the feature vector of the first contact finger pulse wave signal; the back hand vein feature authentication module is used to compare and verify the similarity between the feature vector of the user's second back hand vein image and the feature vector of the first back hand vein image; the dual pulse wave similarity verification module is used to perform similarity analysis on the user's contact finger pulse wave signal and non-contact back hand pulse wave signal to determine whether the dual pulse waves belong to the same authentication subject; and a final identity authentication result is generated based on the joint verification results of the contact finger pulse wave feature authentication module, the back hand vein feature authentication module, and the dual pulse wave similarity verification module.

2. A multimodal identity authentication system based on multiple pulse waves and veins according to claim 1, characterized in that: The contact pulse wave acquisition module includes an infrared pulse wave sensor, which includes an infrared finger clip probe and an acquisition module for acquiring contact finger pulse wave signals.

3. The multimodal identity authentication system based on multiple pulse waves and veins according to claim 1, characterized in that: The hand back video acquisition module includes a near-infrared LED lamp and a near-infrared camera; the near-infrared LED lamp serves as the light source for the near-infrared camera to shoot; the near-infrared camera is used to record the user's hand back video and hand back image, and further obtain the non-contact hand back pulse wave and hand back vein image.

4. The multimodal identity authentication system based on multiple pulse waves and veins according to claim 1, characterized in that: The non-contact hand back pulse wave acquisition module acquires the non-contact hand back pulse wave based on the near-infrared hand back video collected by the hand back video acquisition module, specifically as follows: 1) Frame the near-infrared hand back video; 2) Perform vein detection on each frame of the hand back image to obtain a vein image; 3) averaging the pixels in the vein area of ​​each frame; 4) constructing a set of time-varying sequences based on the pixel averages obtained in step 3) according to the video sequence; 5) The sequence obtained in step 4) is filtered to obtain a pulse wave.

5. A multimodal identity authentication system based on multiple pulse waves and veins according to claim 1, characterized in that: The pulse wave signal preprocessing and feature extraction are specifically as follows: performing wavelet threshold denoising on the collected pulse wave; performing peak detection on the denoised pulse wave signal to segment the single-cycle pulse wave; performing kernel principal component analysis and principal component analysis on the single-cycle pulse wave to complete dimensionality reduction of the pulse wave signal and obtain a feature vector as the first contact finger pulse wave sample.

6. A multimodal identity authentication system based on multiple pulse waves and veins according to claim 1, characterized in that: The similarity comparison between the contact finger pulse wave feature authentication module and the hand back vein feature authentication module is specifically as follows: the pulse wave or hand back vein feature vector X (x1, x2, ..., x2) extracted in the authentication stage is calculated. n ) and the pulse wave or hand vein feature vector Y1(y 11 ,y 12 ,…,y 1n ), Y2…Y m The Euclidean distance of: The k-nearest neighbor classifier is used for classification recognition and similarity comparison.

7. The multimodal identity authentication system based on multiple pulse waves and veins according to claim 1, characterized in that: The similarity analysis of the contact finger pulse wave signal and the non-contact hand back pulse wave signal is specifically as follows: the finger pulse wave signal is recorded as P1, and the hand back pulse wave signal is recorded as P2; P1 and P2 are normalized in amplitude and time, recorded as P1′ and P2′; then the Euclidean distance d(P1′, P2′) between P1′ and P2′ is calculated. If d(P1′, P2′) is less than a threshold, it is determined that P1 and P2 belong to the same subject.

8. A multimodal identity authentication method based on multiple pulse waves and veins, characterized in that: The method is divided into a registration phase and an authentication phase, including the following steps: During the registration phase, S1 uses a near-infrared camera to capture the back of the registered user's hand image to obtain a first back of the hand vein image, and stores the first back of the hand vein image in the database as authentication reference information; S2, during the registration phase, uses a contact finger pulse wave sensor to collect a contact finger pulse wave signal to obtain a first contact finger pulse wave sample, and stores the first contact finger pulse wave sample in a database as authentication reference information; During the authentication phase, S3 uses a near-infrared camera to collect video information of the back of the hand of the authenticated user to obtain a second back-hand vein image and a non-contact back-hand pulse wave signal; In step S4, during the authentication phase, the contact finger pulse wave sensor is used to collect the contact finger pulse wave signal of the authenticated user to obtain a second contact finger pulse wave sample. During the collection of the second contact finger pulse wave sample, the back of the hand video information is continuously obtained through the near-infrared camera. After the contact finger pulse wave signal is collected, step S5 is executed. S5 performs vein recognition authentication by comparing the second hand back vein image obtained in step S3 with the first hand back vein image by similarity; S6: performing contact finger pulse wave authentication by comparing the second contact finger pulse wave sample obtained in step S4 with the first contact finger pulse wave sample through similarity; S7 performs dual pulse wave similarity verification on the contact finger pulse wave signal and the non-contact hand back pulse wave signal obtained in step S4; S8 generates an identity authentication result based on the combined verification results of hand back vein recognition authentication, contact pulse wave authentication and dual pulse wave similarity verification.

9. The multimodal identity authentication method based on multiple pulse waves and veins according to claim 8, characterized in that: The specific process of step S1 is: collecting the back of the hand image information of the registered user, detecting whether the collected image is the back of the hand through the back of the hand detection algorithm, and if so, saving it as the first back of the hand vein image, otherwise re-collecting the image.

Images