Athlete identity verification system and method for sports events

Abstract

The present application relates to the field of identity verification technology, and specifically discloses a system and method for athlete identity verification for sports events, which utilizes palm vein imaging technology to embed the athlete's palm vein image into the athlete's competition certificate during the event registration stage, and collects the athlete's palm vein image through a palm vein recognition device at the event site, and compares and analyzes it with the pre-stored information in the certificate to achieve fast and accurate identity verification. In this way, the limitations of traditional identity verification methods can be effectively overcome, and the security of sports events and the efficiency of identity verification can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of identity verification, and more specifically, to a system and method for verifying the identity of athletes used in sports events. Background Art

[0002] In sports competitions, ensuring the accuracy of the identity of participating athletes is an important part of ensuring fair competition and event management. Traditionally, athletes need to fill in detailed personal information when registering, including name, gender, age, ID number, club or unit, etc. When entering the event, athletes need to bring valid identification documents, such as ID cards, passports, etc., so that the organizing committee can compare with the registration information. However, this verification method that relies on manual verification of identity documents is inefficient and prone to errors. It is difficult to quickly and effectively complete the identity verification of a large number of athletes in large-scale events, and there is a risk of forged documents.

[0003] The existing identity verification system based on fingerprint and facial recognition technology has greatly improved the security and convenience of identity authentication. However, due to the occupation of athletes, fingerprints may be severely worn and fingerprint recognition may not be possible. At the same time, facial recognition technology also has certain limitations and is easily affected by factors such as dressing, facial expression changes and ambient light, which may affect the accuracy of recognition.

[0004] Therefore, an optimized athlete identity verification system and method for sports events is desired. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed. The embodiment of this application provides an athlete identity verification system and method for sports events, which uses palm vein imaging technology to embed the athlete's palm vein image into the athlete's competition certificate during the event registration stage. At the event site, the athlete's palm vein image is collected by a palm vein recognition device and compared with the pre-stored information in the certificate to achieve fast and accurate identity verification. In this way, the limitations of traditional identity verification methods can be effectively overcome, and the security and efficiency of identity verification of sports events can be improved.

[0006] Accordingly, according to one aspect of the present application, there is provided a system for verifying the identity of athletes in sports events, which includes:

[0007] A palm vein image acquisition module, used to acquire a palm vein detection image of the identified object acquired by a palm vein image acquisition device, and extract a palm vein verification image from the competition certificate of the identified object;

[0008] A multi-level feature extraction module, used for performing multi-level feature extraction on the palm vein detection image to obtain a multi-level feature interactive fusion representation vector for detecting palm veins;

[0009] A verification feature generation module, used to generate a verification palm vein multi-level feature interactive fusion representation vector based on the palm vein verification image;

[0010] A difference measurement module is used to extract the feature difference between the detection palm vein multi-level feature interactive fusion representation vector and the verification palm vein multi-level feature interactive fusion representation vector to obtain an identity authentication difference feature vector;

[0011] The identity verification module is used to determine whether the identified object matches the athlete information pre-stored in its competition certificate based on the identity verification difference feature vector.

[0012] In the above-mentioned athlete identity verification system for sports events, the multi-level feature extraction module includes: a texture feature extraction unit, used to extract texture features of the palm vein detection image to obtain a detected palm vein texture feature vector; a global structure semantic feature extraction unit, used to extract global structure semantic features of the palm vein detection image to obtain a detected palm vein structure semantic feature vector; a multi-level feature fusion unit, used to cascade fuse the detected palm vein texture feature vector and the detected palm vein structure semantic feature vector to obtain the detected palm vein multi-level feature interactive fusion representation vector.

[0013] In the above-mentioned athlete identity verification system for sports events, the texture feature extraction unit includes: a grayscale subunit, used to convert the palm vein detection image into a grayscale image to obtain a palm vein detection grayscale image; a texture statistical feature extraction subunit, used to calculate the local binary pattern histogram of the palm vein detection grayscale image to obtain the detected palm vein texture feature vector.

[0014] In the above-mentioned athlete identity verification system for sports events, the global structure semantic feature extraction unit is used to: pass the palm vein detection image through a palm vein global structure feature extractor based on a VIT model to obtain the detected palm vein structure semantic feature vector.

[0015] In the above-mentioned athlete identity verification system for sports events, the difference measurement module is used to: calculate the differential feature vector between the detected palm vein multi-level feature interactive fusion representation vector and the verified palm vein multi-level feature interactive fusion representation vector as the identity verification difference feature vector.

[0016] In the above-mentioned athlete identity verification system for sports events, the identity verification module includes: a feature optimization unit, used to perform a reverse interpretability analysis based on backward learning sensitive preference on the identity verification difference feature vector to obtain an optimized identity verification difference feature vector; a verification result generation unit, used to pass the optimized identity verification difference feature vector through a classifier-based athlete authenticator to obtain a verification result, and the verification result is used to indicate whether the identified object matches the athlete information pre-stored in the competition certificate.

[0017] In the above-mentioned athlete identity verification system for sports events, the feature optimization unit includes: an activation subunit, which is used to perform probabilistic activation on the identity verification difference feature vector based on the Sigmoid function to obtain a probabilistic identity verification difference feature vector; a label probability generation subunit, which is used to input the identity verification difference feature vector into a classifier to obtain an initial class label probability value; an error metric subunit, which is used to calculate the absolute value of the difference between the feature value at each position in the probabilistic identity verification difference feature vector and the initial class label probability value to obtain a backward machine learning fine-grained error metric vector; an attention optimization subunit, which is used to generate a label probability subunit using the backward machine learning fine-grained error metric vector. The quantity vector is used as a key vector, the identity verification difference feature vector is used as a query vector and a value vector, and the key vector, the query vector and the value vector are input into a converter structure to obtain a backward learning type attention optimized identity verification difference feature vector; a scaling factor calculation subunit is used to calculate the ratio between the sum of the backward machine learning fine-grained error metric vector and the length of the backward machine learning fine-grained error metric vector, and calculate the natural exponential function value with the ratio as the power as the optimization scaling factor; an optimization subunit is used to calculate the product between the optimization scaling factor and the backward learning type attention optimized identity verification difference feature vector to obtain the optimized identity verification difference feature vector.

[0018] In the above-mentioned athlete identity verification system for sports events, the verification result generating unit includes: a fully connected encoding unit, used to use the fully connected layer of the athlete identity verification device to fully connect the optimized identity verification difference feature vector to obtain a fully connected encoded feature vector; a probabilization unit, used to pass the fully connected encoded feature vector through the Softmax classification function of the athlete identity verification device to obtain the probability value of the optimized identity verification difference feature vector belonging to each classification label, wherein the classification labels include match and mismatch; a verification result determining unit, used to determine the classification label corresponding to the largest of the probability values ​​as the verification result.

[0019] According to another aspect of the present application, a method for verifying the identity of an athlete in a sports event is provided, comprising:

[0020] Acquire a palm vein detection image of the identified object captured by a palm vein image acquisition device, and extract a palm vein verification image from the competition certificate of the identified object;

[0021] Performing multi-level feature extraction on the palm vein detection image to obtain a multi-level feature interactive fusion representation vector for detecting palm veins;

[0022] Based on the palm vein verification image, generating a verification palm vein multi-level feature interactive fusion representation vector;

[0023] Extracting the feature difference between the detected palm vein multi-level feature interactive fusion representation vector and the verified palm vein multi-level feature interactive fusion representation vector to obtain an identity authentication difference feature vector;

[0024] Based on the identity verification difference feature vector, it is determined whether the identified object matches the athlete information pre-stored in its competition certificate.

[0025] Compared with the prior art, the athlete identity verification system and method provided by the present application for sports events utilizes palm vein imaging technology to embed the athlete's palm vein image into the athlete's competition certificate during the event registration stage. At the event site, the athlete's palm vein image is collected by a palm vein recognition device and compared and analyzed with the pre-stored information in the certificate to achieve fast and accurate identity verification. In this way, the limitations of traditional identity verification methods can be effectively overcome, and the security and efficiency of identity verification in sports events can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0027] Figure 1 The block diagram is a system for verifying the identity of athletes in sports events according to an embodiment of the present application.

[0028] Figure 2 Schematic diagram of the architecture of an athlete identity verification system for sports events according to an embodiment of the present application.

[0029] Figure 3 It is a block diagram of a multi-level feature extraction module in an athlete identity verification system for sports events according to an embodiment of the present application.

[0030] Figure 4 The block diagram is a texture feature extraction unit in a system for verifying the identity of an athlete in a sports event according to an embodiment of the present application.

[0031] Figure 5 The block diagram is a identity verification module in a player identity verification system for sports events according to an embodiment of the present application.

[0032] Figure 6 The present invention is a flowchart of a method for verifying the identity of athletes in sports events according to an embodiment of the present application. DETAILED DESCRIPTION

[0033] Below, the embodiments of the present application will be described in more detail in conjunction with the accompanying drawings, and the above and other purposes, features and advantages of the present application will become more apparent. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein.

[0034] Figure 1 The block diagram is a system for verifying the identity of athletes in sports events according to an embodiment of the present application. Figure 2 FIG. 1 is a schematic diagram of the architecture of an athlete identity verification system for sports events according to an embodiment of the present application. Figure 1 and Figure 2 As shown, according to an embodiment of the present application, the athlete identity verification system 100 for sports events includes: a palm vein image acquisition module 110, which is used to acquire a palm vein detection image of an identified object acquired by a palm vein image acquisition device, and extract a palm vein verification image in the competition certificate of the identified object; a multi-level feature extraction module 120, which is used to perform multi-level feature extraction on the palm vein detection image to obtain a detection palm vein multi-level feature interactive fusion representation vector; a verification feature generation module 130, which is used to generate a verification palm vein multi-level feature interactive fusion representation vector based on the palm vein verification image; a difference measurement module 140, which is used to extract the feature difference between the detection palm vein multi-level feature interactive fusion representation vector and the verification palm vein multi-level feature interactive fusion representation vector to obtain an identity verification difference feature vector; an identity verification module 150, which is used to determine whether the identified object matches the athlete information pre-stored in its competition certificate based on the identity verification difference feature vector.

[0035] As mentioned in the above background technology, the existing identity verification system based on fingerprint and facial recognition technology has greatly improved the security and convenience of identity authentication. However, due to the occupation of athletes, fingerprints may be severely worn and fingerprint recognition cannot be performed. At the same time, facial recognition technology also has certain limitations and is easily affected by factors such as dress, changes in facial expressions and ambient light, which may affect the accuracy of recognition. In response to the above technical problems, the technical concept of this application is to use palm vein imaging technology to embed the athlete's palm vein image into the athlete's competition certificate during the event registration stage. At the event site, the athlete's palm vein image is collected by a palm vein recognition device, and compared and analyzed with the pre-stored information in the certificate to achieve fast and accurate identity authentication. In this way, the limitations of traditional identity verification methods can be effectively overcome, and the security and identity verification efficiency of sports events can be improved.

[0036] In the above-mentioned athlete identity verification system 100 for sports events, the palm vein image acquisition module 110 is used to acquire the palm vein detection image of the identified object collected by the palm vein image acquisition device, and extract the palm vein verification image in the competition certificate of the identified object. It should be understood that the palm vein is a reflection of the internal vascular structure of the human body. The palm vein image of each person is unique and has a high biometric uniqueness. Compared with other biometric technologies (such as fingerprints, facial recognition, etc.), palm vein images are more difficult to forge or copy. The palm vein is located under the skin and is not easily interfered or damaged by external factors. The acquisition process is non-contact, which reduces the possibility of forgery. This feature makes palm vein recognition particularly suitable for high-security occasions such as sports events. In other words, using palm vein images as a means of identity verification can greatly improve the accuracy of identity verification. By comparing the palm vein detection image collected on-site by the athlete with the palm vein verification image pre-stored in the certificate, the authenticity of the athlete's identity can be effectively judged to prevent illegal acts such as impersonation.

[0037] In the above-mentioned athlete identity verification system 100 for sports events, the multi-level feature extraction module 120 is used to perform multi-level feature extraction on the palm vein detection image to obtain a multi-level feature interactive fusion representation vector of the detected palm veins. It should be understood that the palm vein detection image contains local texture features such as branches and intersections of veins, as well as global structural features such as the overall direction and shape of veins. Therefore, in order to improve the robustness of palm vein feature representation and the accuracy of recognition, in the technical solution of the present application, a multi-level feature extraction method is adopted to extract and fuse palm vein image features from different levels and angles, fully explore the identity-related information in the palm vein image, better characterize the complexity of the palm vein image, improve recognition accuracy, and enhance the ability to recognize individual differences.

[0038] Figure 3 FIG. 1 is a block diagram of a multi-level feature extraction module in an athlete identity verification system for sports events according to an embodiment of the present application. Figure 3 As shown, the multi-level feature extraction module 120 includes: a texture feature extraction unit 121, used to extract the texture features of the palm vein detection image to obtain a detected palm vein texture feature vector; a global structure semantic feature extraction unit 122, used to extract the global structure semantic features of the palm vein detection image to obtain a detected palm vein structure semantic feature vector; a multi-level feature fusion unit 123, used to cascade fuse the detected palm vein texture feature vector and the detected palm vein structure semantic feature vector to obtain the detected palm vein multi-level feature interactive fusion representation vector.

[0039] Specifically, the texture feature extraction unit 121 is used to extract the texture features of the palm vein detection image to obtain a texture feature vector for detecting palm veins. Figure 4 FIG. 1 is a block diagram of a texture feature extraction unit in an athlete identity verification system for sports events according to an embodiment of the present application. Figure 4 As shown, the texture feature extraction unit 121 includes: a grayscale subunit 1211, used to convert the palm vein detection image into a grayscale image to obtain a palm vein detection grayscale image; a texture statistical feature extraction subunit 1212, used to calculate the local binary pattern histogram of the palm vein detection grayscale image to obtain the detected palm vein texture feature vector.

[0040] Specifically, the grayscale subunit 1211 is used to convert the palm vein detection image into a grayscale image to obtain a palm vein detection grayscale image. It should be understood that grayscale is the process of converting a color image into a monochrome image, which can reduce the complexity of subsequent processing while retaining the main information of the image. In the technical solution of the present application, by converting the palm vein detection image into a grayscale image, the image data can be simplified, the interference of color information can be avoided, and the texture features of the image can be highlighted, making it easier to analyze and extract.

[0041] Specifically, the texture statistical feature extraction subunit 1212 is used to calculate the local binary pattern histogram of the palm vein detection grayscale image to obtain the detected palm vein texture feature vector. Those skilled in the art should know that Local Binary Patterns (LBP) is a commonly used texture feature description method, which generates a binary code representing local texture information by comparing the grayscale difference between a pixel point in an image and its neighboring pixel points. The LBP operator is sensitive to local details of the texture and has a small amount of calculation, so it is suitable for fast extraction of texture features. In the technical solution of the present application, by calculating the LBP histogram of the palm vein detection grayscale image, normalizing it, and converting it into a vector form, a statistical vector describing the palm vein texture features can be obtained, which effectively captures the local texture changes of the palm vein image and provides key texture feature representation for subsequent identity verification.

[0042] Specifically, the global structure semantic feature extraction unit 122 is used to extract the global structure semantic features of the palm vein detection image to obtain the palm vein structure semantic feature vector. In a specific example of the present application, the processing method for extracting the global structure semantic features of the palm vein detection image is to pass the palm vein detection image through the palm vein global structure feature extractor based on the VIT model to obtain the palm vein structure semantic feature vector. It should be understood that the VIT (Vision Transformer) model is good at processing the global dependency of the image, can understand the complex vein structure at the semantic level, and generate a feature representation that can accurately describe the global structure information of the palm vein. Specifically, the VIT model cuts the input palm vein detection image into multiple small blocks, and processes each small block in parallel through the self-attention mechanism of the Transformer architecture to capture the relationship between each small block and other small blocks, thereby forming global structural information. Through this processing method, global features such as the branching structure, intersection distribution and overall shape of the vein can be effectively extracted, providing more comprehensive structural information support for identity authentication.

[0043] Specifically, the multi-level feature fusion unit 123 is used to cascade fuse the detected palm vein texture feature vector and the detected palm vein structure semantic feature vector to obtain the detected palm vein multi-level feature interactive fusion representation vector. It should be understood that the detected palm vein texture feature vector and the detected palm vein structure semantic feature vector respectively contain the local texture details and global structure information of the palm vein image. These two types of features complement each other at different levels. By comprehensively utilizing the texture and global structure information of the image, richer identity recognition features can be provided. In the technical solution of the present application, the detected palm vein texture feature vector and the detected palm vein structure semantic feature vector are spliced ​​and fused through a cascade operation, which not only retains the characteristics of the original features, but also promotes the mutual complementation between features, enhances the expression ability of features, and makes the model more robust in the face of individual differences and noise interference.

[0044] In the above-mentioned athlete identity verification system 100 for sports events, the verification feature generation module 130 is used to generate a verification palm vein multi-level feature interactive fusion representation vector based on the palm vein verification image. That is, in the same way, the palm vein verification image is processed using the same image processing method as the detection palm vein image, through image graying and calculation of local binary pattern histogram. Based on the global structural semantic feature extraction of the ViT model and the cascade fusion of multi-level image features, a verification palm vein multi-level feature interactive fusion representation vector is generated. In this way, the palm vein features of the athletes are processed consistently, ensuring consistency and fairness in the feature extraction stage, which is conducive to improving the accuracy and reliability of identity verification.

[0045] In the above-mentioned athlete identity verification system 100 for sports events, the difference measurement module 140 is used to extract the feature difference between the multi-level feature interactive fusion representation vector of the detected palm vein and the multi-level feature interactive fusion representation vector of the verified palm vein to obtain the identity verification difference feature vector. In a specific example of the present application, the processing method for extracting the feature difference between the multi-level feature interactive fusion representation vector of the detected palm vein and the multi-level feature interactive fusion representation vector of the verified palm vein is to calculate the differential feature vector between the multi-level feature interactive fusion representation vector of the detected palm vein and the multi-level feature interactive fusion representation vector of the verified palm vein as the identity verification difference feature vector. That is, by calculating the element-by-element difference between the multi-level feature interactive fusion representation vector of the detected palm vein and the multi-level feature interactive fusion representation vector of the verified palm vein to quantify the degree of change of the palm vein feature between the detected image and the verified image, the inconsistency between the feature representation of the two is effectively captured, thereby obtaining the identity verification difference feature vector, providing key difference information for subsequent identity verification decisions.

[0046] In the above-mentioned athlete identity verification system 100 for sports events, the identity verification module 150 is used to determine whether the identified object matches the athlete information pre-stored in its competition certificate based on the identity verification difference feature vector. Figure 5 FIG. 1 is a block diagram of an identity verification module in an athlete identity verification system for sports events according to an embodiment of the present application. Figure 5 As shown, the identity authentication module 150 includes: a feature optimization unit 151, used for performing a reverse interpretability analysis based on backward learning sensitive preference on the identity authentication difference feature vector to obtain an optimized identity authentication difference feature vector; a verification result generation unit 152, used for passing the optimized identity authentication difference feature vector through a classifier-based athlete authenticator to obtain a verification result, and the verification result is used to indicate whether the identified object matches the athlete information pre-stored in the competition certificate.

[0047] Specifically, the feature optimization unit 151 is used to perform a reverse interpretability analysis based on backward learning sensitive preference on the identity authentication difference feature vector to obtain an optimized identity authentication difference feature vector. In particular, in the technical solution of the present application, it is considered that converting the palm vein detection image into a grayscale image will lose color information, and some color information may have a certain influence on the recognition of palm vein texture. And although Local Binary Patterns (LBP) is an effective texture descriptor, it may not be able to fully capture all important texture features, especially when the texture changes are subtle. At the same time, the palm vein global structure feature extractor based on the Vision Transformer (VIT) model may perform poorly on certain specific palm vein structures, resulting in incomplete feature extraction, which in turn causes the identity authentication difference feature vector to have a class harmony offset relative to the predetermined class label of the classifier. Therefore, in order to reduce the class harmony offset of the identity authentication difference feature vector, in the technical solution of the present application, the identity authentication difference feature vector is further subjected to a reverse interpretability analysis based on backward learning sensitive preference.

[0048] Specifically, the feature optimization unit 151 includes: an activation subunit, which is used to perform probabilistic activation on the identity authentication difference feature vector based on the Sigmoid function to obtain a probabilistic identity authentication difference feature vector; a label probability generation subunit, which is used to input the identity authentication difference feature vector into a classifier to obtain an initial class label probability value; an error measurement subunit, which is used to calculate the absolute value of the difference between the feature value of each position in the probabilistic identity authentication difference feature vector and the initial class label probability value to obtain a backward machine learning fine-grained error measurement vector; an attention optimization subunit, which is used to use the backward machine learning fine-grained error measurement vector as a key vector , the identity authentication difference feature vector is used as a query vector and a value vector, and the key vector, the query vector and the value vector are input into a converter structure to obtain a backward learning type attention optimized identity authentication difference feature vector; a scaling factor calculation subunit is used to calculate the ratio between the sum of the backward machine learning fine-grained error metric vector and the length of the backward machine learning fine-grained error metric vector, and calculate the value of a natural exponential function with the ratio as a power as an optimized scaling factor; an optimization subunit is used to calculate the product between the optimized scaling factor and the backward learning type attention optimized identity authentication difference feature vector to obtain the optimized identity authentication difference feature vector.

[0049] That is, the identity authentication difference feature vector is subjected to reverse interpretability analysis based on backward learning sensitive preference using the following feature optimization formula to obtain an optimized identity authentication difference feature vector, wherein the feature optimization formula is: ; ;in, represents a probabilistic identity authentication difference feature vector obtained after probabilistic activation of the identity authentication difference feature vector, represents the initial class label probability value, It means subtracting by position. represents the fine-grained error metric vector for backward machine learning, represents the first The eigenvalues ​​at the positions, represents the length of the backward machine learning fine-grained error metric vector, represents the transpose of a vector, represents vector multiplication, represents the normalized exponential function, It means point multiplication by position. represents the optimized authentication difference feature vector.

[0050] Based on this, in the technical solution of the present application, the identity authentication difference feature vector is subjected to reverse interpretability analysis based on backward learning sensitive preference, which first implements probabilistic activation based on Sigmoid function on the identity authentication difference feature vector to obtain a probabilistic identity authentication difference feature vector, and inputs the identity authentication difference feature vector into the classifier to obtain the initial class label probability value. Then, the error between the probabilistic identity authentication difference feature vector and the initial class label probability value is calculated through backward error analysis to generate a fine-grained error metric vector, which quantifies the viewpoint metric contribution of each feature position to the classification error.

[0051] Then, the backward machine learning fine-grained error metric vector is used as the key vector, the identity verification difference feature vector is used as the query vector and the value vector, and the key vector, the query vector and the value vector are input into the transformer structure to perform error metric-based attention-based backward machine learning on the identity verification difference feature vector through a Transformer-like architecture. In this way, the identity verification difference feature vector is subjected to a backward interpretability analysis so that the obtained backward learning-like attention-optimized identity verification difference feature vector has a backward learning sensitive preference relative to the predetermined class label of the classifier. And further, the ratio of the sum of the error metric vector and the vector length is calculated, and the optimized scaling factor obtained by calculating the natural exponential function value with this ratio as the exponent is used to perform backward scaling optimization on the backward learning-like attention-optimized identity verification difference feature vector. In this way, the sensitive preference and robustness of the identity verification difference feature vector relative to the predetermined class label of the classifier can be effectively improved, thereby improving the accuracy of the classification result.

[0052] Specifically, the verification result generating unit 152 is used to pass the optimized identity authentication difference feature vector through the athlete identity authentication device based on the classifier to obtain a verification result, and the verification result is used to indicate whether the identified object matches the athlete information pre-stored in the competition certificate. It should be understood that in the classifier, by performing multi-layer nonlinear transformation on the optimized identity authentication difference feature vector, the abstract feature pattern in the optimized identity authentication difference feature vector can be learned, the feature difference of the athlete's palm vein feature in the detection image and the verification image can be understood, and a probability distribution can be output to indicate the possibility of the identified object matching the athlete information pre-stored in its competition certificate and the possibility of mismatching. Then, the classification result with the highest probability is selected as the final verification result, so as to achieve accurate verification of the athlete's identity.

[0053] Specifically, the verification result generating unit 152 includes: a fully connected encoding unit, used to use the fully connected layer of the athlete identity authentication device to fully connect the optimized identity authentication difference feature vector to obtain a fully connected encoded feature vector; a probabilization unit, used to pass the fully connected encoded feature vector through the Softmax classification function of the athlete identity authentication device to obtain the probability value of the optimized identity authentication difference feature vector belonging to each classification label, wherein the classification labels include match and mismatch; a verification result determining unit, used to determine the classification label corresponding to the largest probability value as the verification result.

[0054] In summary, according to the embodiment of the present application, the athlete identity verification system for sports events is explained, which uses palm vein imaging technology to embed the athlete's palm vein image into the athlete's competition certificate during the event registration stage. At the event site, the athlete's palm vein image is collected by the palm vein recognition device and compared with the pre-stored information in the certificate to achieve fast and accurate identity verification. In this way, the limitations of traditional identity verification methods can be effectively overcome, and the security and efficiency of identity verification of sports events can be improved.

[0055] Figure 6 FIG. 1 is a flow chart of a method for verifying the identity of athletes in sports events according to an embodiment of the present application. Figure 6 As shown, according to the embodiment of the present application, the method for verifying the identity of athletes in sports events includes the following steps: S110, obtaining a palm vein detection image of an identified object captured by a palm vein image acquisition device, and extracting a palm vein verification image in the competition certificate of the identified object; S120, performing multi-level feature extraction on the palm vein detection image to obtain a multi-level feature interactive fusion representation vector of the detected palm vein; S130, generating a multi-level feature interactive fusion representation vector of the verified palm vein based on the palm vein verification image; S140, extracting the feature difference between the multi-level feature interactive fusion representation vector of the detected palm vein and the multi-level feature interactive fusion representation vector of the verified palm vein to obtain an identity verification difference feature vector; S150, determining whether the identified object matches the athlete information pre-stored in its competition certificate based on the identity verification difference feature vector.

[0056] Here, those skilled in the art will appreciate that the specific operations of each step in the above-mentioned method for verifying the identity of an athlete in a sports event have been described in detail above. Figures 1 to 5 The present invention has been described in detail in the description of the athlete identity verification system for sports events, and therefore, its repeated description will be omitted.

[0057] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0058] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the module division is only a logical function division, and there may be other division methods in actual implementation. The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0059] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0060] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0061] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0062] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0063] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A system for verifying the identity of athletes in sports events, characterized in that: include: A palm vein image acquisition module, used to acquire a palm vein detection image of the identified object acquired by a palm vein image acquisition device, and extract a palm vein verification image from the competition certificate of the identified object; A multi-level feature extraction module, used for performing multi-level feature extraction on the palm vein detection image to obtain a multi-level feature interactive fusion representation vector for detecting palm veins; A verification feature generation module, used to generate a verification palm vein multi-level feature interactive fusion representation vector based on the palm vein verification image; A difference measurement module is used to extract the feature difference between the detection palm vein multi-level feature interactive fusion representation vector and the verification palm vein multi-level feature interactive fusion representation vector to obtain an identity authentication difference feature vector; An identity verification module, used to determine whether the identified object matches the athlete information pre-stored in its competition certificate based on the identity verification difference feature vector; Wherein, the identity verification module includes: A feature optimization unit, configured to perform a reverse interpretability analysis based on a backward learning sensitive preference on the identity authentication difference feature vector to obtain an optimized identity authentication difference feature vector; Wherein, the feature optimization unit comprises: An activation subunit, used for performing probabilistic activation on the identity authentication difference feature vector based on a Sigmoid function to obtain a probabilistic identity authentication difference feature vector; a label probability generating subunit, used for inputting the identity verification difference feature vector into a classifier to obtain an initial class label probability value; An error metric subunit, used to calculate the absolute value of the difference between the feature value at each position in the probabilistic identity authentication difference feature vector and the initial class label probability value to obtain a backward machine learning fine-grained error metric vector; an attention optimization subunit, configured to use the backward machine learning fine-grained error metric vector as a key vector, the identity verification difference feature vector as a query vector and a value vector, and input the key vector, the query vector and the value vector into a converter structure to obtain a backward learning class attention optimized identity verification difference feature vector; A scaling factor calculation subunit, used to calculate the ratio between the sum of the backward machine learning fine-grained error metric vector and the length of the backward machine learning fine-grained error metric vector, and calculate a natural exponential function value with the ratio as a power as an optimization scaling factor; The optimization subunit is used to calculate the product between the optimization scaling factor and the backward learning class attention optimized identity authentication difference feature vector to obtain the optimized identity authentication difference feature vector.

2. The athlete identity verification system for sports events according to claim 1, characterized in that: The multi-level feature extraction module comprises: A texture feature extraction unit, used for extracting texture features of the palm vein detection image to obtain a texture feature vector for detecting palm veins; A global structure semantic feature extraction unit, used for extracting the global structure semantic features of the palm vein detection image to obtain a detected palm vein structure semantic feature vector; The multi-level feature fusion unit is used to cascade fuse the detected palm vein texture feature vector and the detected palm vein structure semantic feature vector to obtain the detected palm vein multi-level feature interactive fusion representation vector.

3. The athlete identity verification system for sports events according to claim 2, characterized in that: The texture feature extraction unit comprises: A grayscale subunit, used for converting the palm vein detection image into a grayscale image to obtain a palm vein detection grayscale image; The texture statistical feature extraction subunit is used to calculate the local binary pattern histogram of the palm vein detection grayscale image to obtain the detected palm vein texture feature vector.

4. The athlete identity verification system for sports events according to claim 3, characterized in that: The global structure semantic feature extraction unit is used to: The palm vein detection image is passed through a palm vein global structure feature extractor based on a VIT model to obtain a semantic feature vector of the detected palm vein structure.

5. The athlete identity verification system for sports events according to claim 4, characterized in that: The difference measurement module is used to: A differential feature vector between the detected palm vein multi-level feature interactive fusion representation vector and the verified palm vein multi-level feature interactive fusion representation vector is calculated as the identity verification difference feature vector.

6. The athlete identity verification system for sports events according to claim 5, characterized in that: The identity verification module further includes: The verification result generating unit is used to pass the optimized identity verification difference feature vector through a classifier-based athlete identity verification device to obtain a verification result, wherein the verification result is used to indicate whether the identified object matches the athlete information pre-stored in the competition certificate.

7. The athlete identity verification system for sports events according to claim 6, characterized in that: The verification result generating unit comprises: a fully connected encoding unit, configured to perform fully connected encoding on the optimized identity verification difference feature vector using the fully connected layer of the athlete identity verification device to obtain a fully connected encoded feature vector; A probabilization unit, configured to pass the fully connected encoded feature vector through a Softmax classification function of the athlete identity verification device to obtain a probability value of the optimized identity verification difference feature vector belonging to each classification label, wherein the classification labels include match and mismatch; A verification result determination unit is used to determine the classification label corresponding to the largest one of the probability values ​​as the verification result.

8. A method for verifying the identity of an athlete in a sports event, using the system for verifying the identity of an athlete in a sports event as claimed in claim 1, characterized in that: include: Acquire a palm vein detection image of the identified object captured by a palm vein image acquisition device, and extract a palm vein verification image from the competition certificate of the identified object; Performing multi-level feature extraction on the palm vein detection image to obtain a multi-level feature interactive fusion representation vector for detecting palm veins; Based on the palm vein verification image, generating a verification palm vein multi-level feature interactive fusion representation vector; Extracting the feature difference between the detected palm vein multi-level feature interactive fusion representation vector and the verified palm vein multi-level feature interactive fusion representation vector to obtain an identity authentication difference feature vector; Based on the identity verification difference feature vector, it is determined whether the identified object matches the athlete information pre-stored in its competition certificate.

9. The method for verifying the identity of athletes in sports events according to claim 8, characterized in that: Performing multi-level feature extraction on the palm vein detection image to obtain a multi-level feature interactive fusion representation vector for detecting palm veins, including: Extracting texture features of the palm vein detection image to obtain a detected palm vein texture feature vector; Extracting global structural semantic features of the palm vein detection image to obtain a detected palm vein structural semantic feature vector; The detected palm vein texture feature vector and the detected palm vein structure semantic feature vector are cascade-fused to obtain the detected palm vein multi-level feature interactive fusion representation vector.

Images