Iris recognition intelligent chip performance optimization system based on heterogeneous computing technology

Abstract

The invention relates to the technical field of chip performance optimization, in particular to an iris recognition intelligent chip performance optimization system based on a heterogeneous computing technology. Comprising a front-end signal processing module, a feature extraction module, a heterogeneous calculation scheduling module, a feature matching authentication module, an anti-counterfeiting detection module, an energy efficiency optimization module, a privacy protection encryption module and an adaptive learning updating module. The front-end signal processing module is used for image preprocessing and noise filtering; the advantages of different computing resources are fully utilized through heterogeneous computing scheduling, efficient parallel computing is achieved, the iris recognition process is accelerated, cheating through high-resolution pictures or imitations is effectively prevented through anti-counterfeiting detection, and the safety and reliability of the system are improved; the self-adaptive learning updating enables the system to continuously learn and adapt to the change of the iris of the user and the change of the environment, and the long-term effectiveness of recognition is kept.

Description

Technical Field

[0001] The present invention relates to the technical field of chip performance optimization, and in particular to an iris recognition intelligent chip performance optimization system based on heterogeneous computing technology. Background Art

[0002] Iris recognition is a biometric technology that uses the unique features of the iris (the colored ring-shaped part of the eye) to identify individuals. Each person's iris texture is unique, which makes iris recognition a very accurate and secure identity authentication method. An iris recognition smart chip is a dedicated hardware chip that integrates an iris recognition algorithm. This chip can efficiently perform iris image processing and feature matching, thereby achieving fast and accurate identity authentication. Iris recognition smart chips are widely used in security access control, mobile payment, personal identity authentication and other fields.

[0003] The main reasons for optimizing the performance of iris recognition smart chips include improving recognition accuracy. By optimizing algorithms and hardware design, the false recognition rate and rejection rate can be reduced, and recognition accuracy can be improved. The optimized chip can process iris images and perform feature matching faster, thereby shortening recognition time and improving user experience. Energy efficiency optimization technology can reduce chip power consumption, extend the life of the device, and reduce operating costs. The optimized chip can provide stronger data encryption and privacy protection functions to ensure the security of user information.

[0004] However, in general, traditional performance optimization solutions often rely on only a single computing resource (such as CPU), and cannot fully utilize the acceleration capabilities of other computing resources (such as GPU, FPGA). They may be insufficient in anti-counterfeiting detection and are easily deceived by high-resolution images or imitations. They cannot adapt to changes in the user's iris and the environment, resulting in a decrease in recognition accuracy over time. They consume a lot of energy in high-performance mode, and their performance cannot meet the requirements in low-power mode. Finally, traditional solutions have loopholes in data encryption and privacy protection, and are vulnerable to attacks and data leakage risks.

[0005] In summary, it is necessary to propose an iris recognition intelligent chip performance optimization system based on heterogeneous computing technology to solve the above problems. Summary of the invention

[0006] The purpose of the present invention is to provide an iris recognition smart chip performance optimization system based on heterogeneous computing technology to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] The iris recognition smart chip performance optimization system based on heterogeneous computing technology includes front-end signal processing module, feature extraction module, heterogeneous computing scheduling module, feature matching authentication module, anti-counterfeiting detection module, energy efficiency optimization module, privacy protection encryption module, and adaptive learning update module;

[0009] The front-end signal processing module is used for image preprocessing and noise filtering;

[0010] The feature extraction module is used for iris segmentation and texture feature extraction;

[0011] The heterogeneous computing scheduling module is used for resource allocation and parallel computing optimization;

[0012] The feature matching authentication module is used for template matching and dynamic threshold adjustment;

[0013] The anti-counterfeiting detection module is used for deep learning forgery detection and iris liveness detection;

[0014] The energy efficiency optimization module is used for dynamic voltage and frequency regulation and low power consumption mode management;

[0015] The privacy protection encryption module is used for data encryption and anonymous authentication;

[0016] The adaptive learning update module is used for real-time adaptive learning and environmental adaptive adjustment.

[0017] Preferably, the front-end signal processing module further includes an image preprocessing unit and a noise filtering unit;

[0018] The image preprocessing unit uses a multi-scale Retinex algorithm to perform illumination compensation to eliminate the influence of uneven illumination, so as to enhance the contrast and clarity of the iris image;

[0019] The noise filtering unit is used to remove salt and pepper noise in the image by using an adaptive median filter and retain edge details.

[0020] Preferably, the feature extraction module further includes an iris segmentation unit and a texture feature extraction unit;

[0021] The iris segmentation unit is used to identify and segment the iris area and remove interference from the pupil and eyelid by using an edge detection algorithm based on Hough transform;

[0022] The texture feature extraction unit is used to extract the texture features of the iris image by using wavelet transform, and perform multi-scale and multi-directional analysis on the texture features to obtain more discriminative features.

[0023] Preferably, the heterogeneous computing scheduling module further includes a resource allocation unit and a parallel computing optimization unit;

[0024] The resource allocation unit is used to allocate tasks among different computing resources, including CPU, GPU, and FPGA, through a dynamic load balancing algorithm to ensure efficient computing;

[0025] The parallel computing optimization unit is used to parallelize the iris recognition algorithm using OpenCL, CUDA framework or dedicated hardware architecture to accelerate the feature extraction and matching process on a multi-core processor or GPU.

[0026] Preferably, the feature matching authentication module further includes a template matching unit and a dynamic threshold adjustment unit;

[0027] The template matching unit is used to compare the templates in the extracted iris feature database using a similarity algorithm to calculate the similarity;

[0028] The dynamic threshold adjustment unit dynamically adjusts the matching threshold using a fuzzy logic algorithm, so as to adaptively optimize the matching result according to the ambient lighting and individual differences of users.

[0029] Preferably, the anti-counterfeiting detection module further includes an iris liveness detection unit and a deep learning counterfeiting detection unit;

[0030] The iris liveness detection unit is used to detect the liveness characteristics of the iris by combining a spectrum analysis algorithm and a pulsation detection algorithm to prevent the system from being deceived by using high-resolution images or imitations;

[0031] The deep learning forgery detection unit trains a convolutional neural network (CNN) model to identify complex forged iris samples and improve the system's anti-attack capability.

[0032] Preferably, the energy efficiency optimization module further includes a dynamic voltage frequency scaling (DVFS) unit and a low power consumption mode management unit;

[0033] The dynamic voltage and frequency scaling (DVFS) unit applies an adaptive energy management algorithm to adjust the operating voltage and frequency of the chip in real time according to the current load to reduce energy consumption;

[0034] The low power mode management unit uses a deep sleep mode algorithm to reduce chip activity during identified idle periods, further reducing energy consumption.

[0035] Preferably, the privacy protection encryption module further includes a data encryption unit and an anonymous authentication unit;

[0036] The data encryption unit uses an elliptic curve cryptography (ECC) algorithm to encrypt the iris feature data to prevent data leakage and theft;

[0037] The anonymous authentication unit protects user privacy through a zero-knowledge proof algorithm, allowing the user to authenticate without revealing his or her true identity.

[0038] Preferably, the adaptive learning update module further comprises a real-time adaptive learning unit and an environment adaptive adjustment unit;

[0039] The real-time adaptive learning unit is used to utilize an online machine learning algorithm, and the system automatically updates the model after each recognition to adapt to the natural changes of the user's iris over time;

[0040] The environment adaptive adjustment unit is used to optimize the recognition accuracy by enabling the system to adapt to environmental changes, including lighting conditions and camera angles, through a reinforcement learning algorithm.

[0041] Preferably, the workflow of the system includes the following steps:

[0042] S1. Image preprocessing: Obtain the original iris image, perform image acquisition, ensure that the image can be presented clearly under different lighting and environmental conditions, perform illumination compensation on the acquired iris image to eliminate the influence of uneven lighting, enhance the contrast and clarity of the image, remove noise in the image, including salt and pepper noise, through noise filtering technology, retain important edge details, and provide high-quality images for subsequent processing;

[0043] S2. Feature extraction: Segment the preprocessed iris image, identify and separate the iris area, exclude the pupil and non-iris parts of the eyelid, and use advanced texture analysis technology to extract key texture features from the iris area to ensure that the information obtained is highly distinguishable for subsequent matching and authentication;

[0044] S3. Heterogeneous computing scheduling: Evaluate the complexity of the current computing task and dynamically distribute the load according to the available computing resources on the chip (including CPU, GPU, FPGA, ASIC) to optimize the utilization of computing resources. Use parallel computing technology to parallelize the iris recognition algorithm on heterogeneous computing resources, improve the speed of feature extraction and matching, and shorten the recognition time.

[0045] S4. Feature matching and authentication: Compare the extracted iris features with the pre-stored database templates, and determine the matching degree by calculating the similarity between the features. According to the current ambient lighting conditions and individual differences, dynamically adjust the matching threshold to improve the accuracy and robustness of recognition;

[0046] S5. Anti-counterfeiting detection: During the iris recognition process, detect whether the iris is alive, ensure that the input image comes from a real iris rather than a high-resolution picture or imitation, apply deep learning technology to identify complex forged iris samples, and improve the system's ability to prevent spoofing attacks;

[0047] S6. Energy efficiency optimization: Dynamically adjust the chip's operating voltage and frequency according to the current computing load to ensure the best balance between performance and energy consumption, reduce unnecessary energy consumption, and automatically switch to low-power mode when the iris recognition system is idle, further reducing the chip's energy consumption and extending the device's working life;

[0048] S7. Privacy protection and encryption: During data transmission and storage, iris feature data is encrypted to prevent unauthorized access and data leakage, ensure the security of user privacy, and achieve anonymous authentication. Users can perform iris recognition without revealing their true identity, thus increasing privacy protection.

[0049] S8. Adaptive learning and updating: After each recognition, the system automatically learns and updates the model to cope with the natural changes of the user's iris over time and maintain the long-term effectiveness of the recognition system. The system automatically adjusts the recognition parameters according to changes in the environment (including lighting and camera angle) to optimize the accuracy and efficiency of recognition.

[0050] S9. System integration and optimization: Integrate all step modules into the iris recognition smart chip to ensure that all parts work in coordination and achieve an efficient process from image acquisition to final recognition. Regularly perform performance testing and optimization on the entire system to ensure its stability and efficiency in different environments and continuously improve the user experience.

[0051] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention fully utilizes the advantages of different computing resources through heterogeneous computing scheduling, realizes efficient parallel computing, accelerates the iris recognition process, and anti-counterfeiting detection effectively prevents deception using high-resolution images or imitations, thereby improving the security and reliability of the system. Adaptive learning updates enable the system to continuously learn and adapt to changes in the user's iris and environmental changes to maintain the long-term effectiveness of recognition. Energy efficiency optimization reduces energy consumption while ensuring performance and extends the service life of the device. Privacy protection encryption ensures the security and privacy of user iris data and prevents data leakage and theft. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a topology diagram of the iris recognition smart chip performance optimization system based on heterogeneous computing technology of the present invention;

[0053] Figure 2 This is a flow chart of the performance optimization method of the iris recognition smart chip based on heterogeneous computing technology of the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Example 1

[0056] See also Figure 1 , the present invention proposes an iris recognition smart chip performance optimization system based on heterogeneous computing technology, including a front-end signal processing module, a feature extraction module, a heterogeneous computing scheduling module, a feature matching authentication module, an anti-counterfeiting detection module, an energy efficiency optimization module, a privacy protection encryption module, and an adaptive learning update module;

[0057] Among them, this system performs image preprocessing and noise filtering through the front-end signal processing module, this system performs iris segmentation and texture feature extraction through the feature extraction module, this system performs resource allocation and parallel computing optimization through the heterogeneous computing scheduling module, this system performs template matching and dynamic threshold adjustment through the feature matching authentication module, this system performs deep learning forgery detection and iris liveness detection through the anti-counterfeiting detection module, this system performs dynamic voltage and frequency adjustment and low power consumption mode management through the energy efficiency optimization module, this system performs data encryption and anonymous authentication through the privacy protection encryption module, and this system performs real-time adaptive learning and environment adaptive adjustment through the adaptive learning update module.

[0058] In this embodiment, it should also be noted that the front-end signal processing module in this system also includes an image preprocessing unit and a noise filtering unit. The image preprocessing unit uses a multi-scale Retinex algorithm to perform illumination compensation to eliminate the impact of uneven illumination and is used to enhance the contrast and clarity of the iris image. The noise filtering unit is used to use an adaptive median filter to remove salt and pepper noise in the image and retain edge details.

[0059] In this embodiment, it should also be noted that the feature extraction module in this system also includes an iris segmentation unit and a texture feature extraction unit. The iris segmentation unit is used to identify and segment the iris area and remove interference from the pupil and eyelid through an edge detection algorithm based on Hough transform. The texture feature extraction unit is used to extract texture features of the iris image using wavelet transform and perform multi-scale and multi-directional analysis on it to obtain more discriminative features.

[0060] In this embodiment, it should also be noted that the heterogeneous computing scheduling module in this system also includes a resource allocation unit and a parallel computing optimization unit. The resource allocation unit is used to allocate tasks among different computing resources, including CPU, GPU, and FPGA, through a dynamic load balancing algorithm to ensure efficient computing. The parallel computing optimization unit is used to parallelize the iris recognition algorithm using the OpenCL or CUDA framework to accelerate the feature extraction and matching process on a multi-core processor or GPU.

[0061] In this embodiment, it should also be noted that the feature matching authentication module in this system also includes a template matching unit and a dynamic threshold adjustment unit. The template matching unit is used to compare the templates in the extracted iris feature database using a similarity algorithm to calculate the similarity. The dynamic threshold adjustment unit uses a fuzzy logic algorithm to dynamically adjust the matching threshold, which is used to adaptively optimize the matching results according to the ambient lighting and individual differences of users.

[0062] In this embodiment, it should also be noted that the anti-counterfeiting detection module in the system also includes an iris liveness detection unit and a deep learning forgery detection unit. The iris liveness detection unit is used to combine the spectral analysis algorithm and the pulsation detection algorithm to detect the liveness characteristics of the iris to prevent the use of high-resolution images or imitations to deceive the system. The deep learning forgery detection unit trains a convolutional neural network (CNN) model to identify complex forged iris samples and enhance the system's anti-attack capabilities.

[0063] In this embodiment, it should also be noted that the energy efficiency optimization module in the system also includes a dynamic voltage and frequency adjustment (DVFS) unit and a low power mode management unit. The dynamic voltage and frequency adjustment (DVFS) unit applies an adaptive energy management algorithm to adjust the operating voltage and frequency of the chip in real time according to the current load to reduce energy consumption. The low power mode management unit adopts a deep sleep mode algorithm to reduce chip activity when identifying idle periods, thereby further reducing energy consumption.

[0064] In this embodiment, it should also be noted that the privacy protection encryption module in this system also includes a data encryption unit and an anonymous authentication unit. The data encryption unit uses an elliptic curve cryptography (ECC) algorithm to encrypt iris feature data to prevent data leakage and theft. The anonymous authentication unit uses a zero-knowledge proof algorithm to protect user privacy, allowing users to authenticate without revealing their true identity.

[0065] In this embodiment, it should also be noted that the adaptive learning update module in the system also includes a real-time adaptive learning unit and an environmental adaptive adjustment unit. The real-time adaptive learning unit is used to utilize an online machine learning algorithm. The system automatically updates the model after each recognition to adapt to the natural changes of the user's iris over time. The environmental adaptive adjustment unit is used to optimize the recognition accuracy by using a reinforcement learning algorithm to enable the system to adapt to environmental changes, including lighting conditions and camera angles.

[0066] See also Figure 2 The workflow of the iris recognition smart chip performance optimization system based on heterogeneous computing technology includes the following steps:

[0067] S1. Image preprocessing: Obtain the original iris image, perform image acquisition, ensure that the image can be presented clearly under different lighting and environmental conditions, perform illumination compensation on the acquired iris image to eliminate the influence of uneven lighting, enhance the contrast and clarity of the image, remove noise in the image, including salt and pepper noise, through noise filtering technology, retain important edge details, and provide high-quality images for subsequent processing;

[0068] S2. Feature extraction: Segment the preprocessed iris image, identify and separate the iris area, exclude the pupil and non-iris parts of the eyelid, and use advanced texture analysis technology to extract key texture features from the iris area to ensure that the information obtained is highly distinguishable for subsequent matching and authentication;

[0069] S3. Heterogeneous computing scheduling: Evaluate the complexity of the current computing task and dynamically distribute the load according to the available computing resources on the chip (including CPU, GPU, FPGA, ASIC) to optimize the utilization of computing resources. Use parallel computing technology to parallelize the iris recognition algorithm on heterogeneous computing resources, improve the speed of feature extraction and matching, and shorten the recognition time.

[0070] S4. Feature matching and authentication: Compare the extracted iris features with the pre-stored database templates, and determine the matching degree by calculating the similarity between the features. According to the current ambient lighting conditions and individual differences, dynamically adjust the matching threshold to improve the accuracy and robustness of recognition;

[0071] S5. Anti-counterfeiting detection: During the iris recognition process, detect whether the iris is alive, ensure that the input image comes from a real iris rather than a high-resolution picture or imitation, apply deep learning technology to identify complex forged iris samples, and improve the system's ability to prevent spoofing attacks;

[0072] S6. Energy efficiency optimization: Dynamically adjust the chip's operating voltage and frequency according to the current computing load to ensure the best balance between performance and energy consumption, reduce unnecessary energy consumption, and automatically switch to low-power mode when the iris recognition system is idle, further reducing the chip's energy consumption and extending the device's working life;

[0073] S7. Privacy protection and encryption: During data transmission and storage, iris feature data is encrypted to prevent unauthorized access and data leakage, ensure the security of user privacy, and achieve anonymous authentication. Users can perform iris recognition without revealing their true identity, thus increasing privacy protection.

[0074] S8. Adaptive learning and updating: After each recognition, the system automatically learns and updates the model to cope with the natural changes of the user's iris over time and maintain the long-term effectiveness of the recognition system. The system automatically adjusts the recognition parameters according to changes in the environment (including lighting and camera angle) to optimize the accuracy and efficiency of recognition.

[0075] S9. System integration and optimization: Integrate all step modules into the iris recognition smart chip to ensure that all parts work in coordination and achieve an efficient process from image acquisition to final recognition. Regularly perform performance testing and optimization on the entire system to ensure its stability and efficiency in different environments and continuously improve the user experience.

[0076] Example 2

[0077] The above system is used in the actual application of iris recognition smart chip performance optimization, specifically, including the following steps:

[0078] S1. Image Preprocessing

[0079] S1.1. Image acquisition: Use a high-resolution camera to capture iris images to ensure image quality;

[0080] S1.2. Illumination compensation: The multi-scale Retinex algorithm is used for illumination compensation to eliminate the influence of uneven illumination, as shown in formula (1):

[0081] R(x,y)=log(I(x,y))-log[F(x,y)*I(x,y)](1);

[0082] Where I(x,y) is the input image, F(x,y) is the Gaussian surround function, and R(x,y) is the image after illumination compensation;

[0083] S1.3. Noise filtering: Use an adaptive median filter to remove salt and pepper noise in the image, see formula (2):

[0084] g(x,y)=median{f(s,t)}(2);

[0085] Where f(s,t) is the pixel value in the filter window, and g(x,y) is the pixel value after filtering;

[0086] S2. Feature extraction

[0087] S2.1. Iris segmentation: Identify and segment the iris region using an edge detection algorithm based on Hough transform, see formula (3):

[0088]

[0089] Where ρ(θ) is the edge point in polar coordinates, and H(x,y) is the result of Hough transform;

[0090] S2.2. Texture feature extraction: The texture features of the iris image are extracted using wavelet transform, see formula (4):

[0091]

[0092] Where f(t) is the input signal, ψ is the wavelet basis function, and W(a,b) is the wavelet transform coefficient;

[0093] S2.3. Feature analysis: multi-scale and multi-directional analysis of the extracted texture features;

[0094] S3. Heterogeneous computing scheduling

[0095] S3.1.Task evaluation: Evaluate the complexity of the current computing task;

[0096] S3.2. Resource allocation: Allocate tasks among different computing resources through dynamic load balancing algorithms;

[0097] S3.3. Parallel computing: Use OpenCL or CUDA framework to parallelize the iris recognition algorithm;

[0098] S4. Feature matching and authentication

[0099] S4.1. Template comparison: Use similarity algorithm to compare the extracted iris features with the templates in the database;

[0100] S4.2. Similarity calculation: Calculate the similarity between features;

[0101] S4.3.Threshold adjustment: Use fuzzy logic algorithm to dynamically adjust the matching threshold;

[0102] S5. Anti-counterfeiting detection

[0103] S5.1. Liveness detection: Detect the liveness characteristics of the iris by combining the spectral analysis algorithm and the pulsation detection algorithm;

[0104] S5.2. Deep learning forgery detection: Train a convolutional neural network (CNN) model to identify forged iris samples;

[0105] S5.3. Forged sample identification: Apply the trained CNN model to detect forged samples on the input image;

[0106] S6. Energy efficiency optimization

[0107] S6.1. Voltage and frequency regulation: Apply adaptive energy management algorithm to dynamically adjust the chip's operating voltage and frequency;

[0108] S6.2. Low power mode management: Reduce chip activity during idle periods and use deep sleep mode algorithms;

[0109] S6.3. Energy consumption monitoring: monitor the energy consumption of the chip in real time and adjust it as needed;

[0110] S7. Privacy protection and encryption

[0111] S7.1. Data encryption: Use the elliptic curve cryptography (ECC) algorithm to encrypt the iris feature data, see formula (5):

[0112] (x1,y1)=k·(x2,y2) (5);

[0113] Where (x1, y1) and (x2, y2) are points on the elliptic curve, and k is the private key;

[0114] S7.2. Anonymous authentication: protect user privacy through zero-knowledge proof algorithm;

[0115] S7.3. Secure storage: Ensure the security of encrypted data during storage and transmission;

[0116] S8. Adaptive learning and updating

[0117] S8.1. Real-time learning: Using online machine learning algorithms, the system automatically updates the model after each recognition;

[0118] S8.2. Environmental adaptation: Through reinforcement learning algorithms, the system adapts to environmental changes;

[0119] S8.3. Model optimization: Regularly evaluate and optimize the performance of the entire recognition model;

[0120] S9.System Integration and Optimization

[0121] S9.1. Module integration: Integrate all step modules into the iris recognition smart chip;

[0122] S9.2. Performance testing: Regularly perform performance testing on the entire system;

[0123] S9.3. System optimization: Optimize the system based on performance test results to improve user experience.

[0124] In summary, through the above steps, this system has the following advantages:

[0125] Efficient computing: Through the heterogeneous computing scheduling module, the advantages of different computing resources can be fully utilized to achieve efficient parallel computing and accelerate the iris recognition process;

[0126] Strong anti-interference ability: The anti-counterfeiting detection module can effectively prevent the use of high-resolution images or imitations for deception, improving the security and reliability of the system;

[0127] Strong adaptability: The adaptive learning update module enables the system to continuously learn and adapt to changes in the user's iris and the environment, maintaining the long-term effectiveness of recognition;

[0128] High energy efficiency ratio: The energy efficiency optimization module can reduce energy consumption while ensuring performance and extend the service life of the equipment;

[0129] Perfect privacy protection: The privacy protection encryption module can ensure the security and privacy of user iris data and prevent data leakage and theft.

[0130] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The iris recognition intelligent chip performance optimization system based on heterogeneous computing technology is characterized by: It includes front-end signal processing module, feature extraction module, heterogeneous computing scheduling module, feature matching authentication module, anti-counterfeiting detection module, energy efficiency optimization module, privacy protection encryption module, and adaptive learning update module; The front-end signal processing module is used for image preprocessing and noise filtering; The feature extraction module is used for iris segmentation and texture feature extraction; The heterogeneous computing scheduling module is used for resource allocation and parallel computing optimization; The feature matching authentication module is used for template matching and dynamic threshold adjustment; The anti-counterfeiting detection module is used for deep learning forgery detection and iris liveness detection; The energy efficiency optimization module is used for dynamic voltage and frequency regulation and low power consumption mode management; The privacy protection encryption module is used for data encryption and anonymous authentication; The adaptive learning update module is used for real-time adaptive learning and environmental adaptive adjustment.

2. The iris recognition smart chip performance optimization system based on heterogeneous computing technology according to claim 1 is characterized by: The front-end signal processing module also includes an image preprocessing unit and a noise filtering unit; The image preprocessing unit uses a multi-scale Retinex algorithm to perform illumination compensation to eliminate the influence of uneven illumination, so as to enhance the contrast and clarity of the iris image; The noise filtering unit is used to remove salt and pepper noise in the image by using an adaptive median filter and retain edge details.

3. The iris recognition smart chip performance optimization system based on heterogeneous computing technology according to claim 2 is characterized by: The feature extraction module also includes an iris segmentation unit and a texture feature extraction unit; The iris segmentation unit is used to identify and segment the iris area and remove interference from the pupil and eyelid by using an edge detection algorithm based on Hough transform; The texture feature extraction unit is used to extract the texture features of the iris image by using wavelet transform, and perform multi-scale and multi-directional analysis on the texture features to obtain more discriminative features.

4. The iris recognition smart chip performance optimization system based on heterogeneous computing technology according to claim 3 is characterized by: The heterogeneous computing scheduling module also includes a resource allocation unit and a parallel computing optimization unit; The resource allocation unit is used to allocate tasks among different computing resources, including CPU, GPU, and FPGA, through a dynamic load balancing algorithm to ensure efficient computing; The parallel computing optimization unit is used to parallelize the iris recognition algorithm using OpenCL, CUDA framework or dedicated hardware architecture to accelerate the feature extraction and matching process on a multi-core processor or GPU.

5. The iris recognition smart chip performance optimization system based on heterogeneous computing technology according to claim 4 is characterized by: The feature matching authentication module also includes a template matching unit and a dynamic threshold adjustment unit; The template matching unit is used to compare the templates in the extracted iris feature database using a similarity algorithm to calculate the similarity; The dynamic threshold adjustment unit dynamically adjusts the matching threshold using a fuzzy logic algorithm, so as to adaptively optimize the matching result according to the ambient lighting and individual differences of users.

6. The iris recognition smart chip performance optimization system based on heterogeneous computing technology according to claim 5 is characterized by: The anti-counterfeiting detection module also includes an iris liveness detection unit and a deep learning counterfeiting detection unit; The iris liveness detection unit is used to detect the liveness characteristics of the iris by combining a spectrum analysis algorithm and a pulsation detection algorithm to prevent the system from being deceived by using high-resolution images or imitations; The deep learning forgery detection unit trains a convolutional neural network (CNN) model to identify complex forged iris samples and improve the system's anti-attack capability.

7. The iris recognition smart chip performance optimization system based on heterogeneous computing technology according to claim 6 is characterized by: The energy efficiency optimization module also includes a dynamic voltage frequency scaling (DVFS) unit and a low power mode management unit; The dynamic voltage and frequency scaling (DVFS) unit applies an adaptive energy management algorithm to adjust the operating voltage and frequency of the chip in real time according to the current load to reduce energy consumption; The low power mode management unit uses a deep sleep mode algorithm to reduce chip activity during identified idle periods, further reducing energy consumption.

8. The iris recognition smart chip performance optimization system based on heterogeneous computing technology according to claim 7 is characterized by: The privacy protection encryption module also includes a data encryption unit and an anonymous authentication unit; The data encryption unit uses an elliptic curve cryptography (ECC) algorithm to encrypt the iris feature data to prevent data leakage and theft; The anonymous authentication unit protects user privacy through a zero-knowledge proof algorithm, allowing the user to authenticate without revealing his or her true identity.

9. The iris recognition smart chip performance optimization system based on heterogeneous computing technology according to claim 8 is characterized by: The adaptive learning update module also includes a real-time adaptive learning unit and an environment adaptive adjustment unit; The real-time adaptive learning unit is used to utilize an online machine learning algorithm, and the system automatically updates the model after each recognition to adapt to the natural changes of the user's iris over time; The environment adaptive adjustment unit is used to optimize the recognition accuracy by enabling the system to adapt to environmental changes, including lighting conditions and camera angles, through a reinforcement learning algorithm.

10. The iris recognition intelligent chip performance optimization system based on heterogeneous computing technology according to claim 9 is characterized in that: The workflow of the system includes the following steps: S1. Image preprocessing: Obtain the original iris image, perform image acquisition, ensure that the image can be presented clearly under different lighting and environmental conditions, perform illumination compensation on the acquired iris image to eliminate the influence of uneven lighting, enhance the contrast and clarity of the image, remove noise in the image, including salt and pepper noise, through noise filtering technology, retain important edge details, and provide high-quality images for subsequent processing; S2. Feature extraction: Segment the preprocessed iris image, identify and separate the iris area, exclude the pupil and non-iris parts of the eyelid, and use advanced texture analysis technology to extract key texture features from the iris area to ensure that the information obtained is highly distinguishable for subsequent matching and authentication; S3. Heterogeneous computing scheduling: Evaluate the complexity of the current computing task and dynamically distribute the load according to the available computing resources on the chip (including CPU, GPU, FPGA, ASIC) to optimize the utilization of computing resources. Use parallel computing technology to parallelize the iris recognition algorithm on heterogeneous computing resources, improve the speed of feature extraction and matching, and shorten the recognition time. S4. Feature matching and authentication: Compare the extracted iris features with the pre-stored database templates, and determine the matching degree by calculating the similarity between the features. According to the current ambient lighting conditions and individual differences, dynamically adjust the matching threshold to improve the accuracy and robustness of recognition; S5. Anti-counterfeiting detection: During the iris recognition process, detect whether the iris is alive, ensure that the input image comes from a real iris rather than a high-resolution picture or imitation, apply deep learning technology to identify complex forged iris samples, and improve the system's ability to prevent spoofing attacks; S6. Energy efficiency optimization: Dynamically adjust the chip's operating voltage and frequency according to the current computing load to ensure the best balance between performance and energy consumption, reduce unnecessary energy consumption, and automatically switch to low-power mode when the iris recognition system is idle, further reducing the chip's energy consumption and extending the device's working life; S7. Privacy protection and encryption: During data transmission and storage, iris feature data is encrypted to prevent unauthorized access and data leakage, ensure the security of user privacy, and achieve anonymous authentication. Users can perform iris recognition without revealing their true identity, thus increasing privacy protection. S8. Adaptive learning and updating: After each recognition, the system automatically learns and updates the model to cope with the natural changes of the user's iris over time and maintain the long-term effectiveness of the recognition system. The system automatically adjusts the recognition parameters according to changes in the environment (including lighting and camera angle) to optimize the accuracy and efficiency of recognition. S9. System integration and optimization: Integrate all step modules into the iris recognition smart chip to ensure that all parts work in coordination and achieve an efficient process from image acquisition to final recognition. Regularly perform performance testing and optimization on the entire system to ensure its stability and efficiency in different environments and continuously improve the user experience.

Images