Medical beauty AIGC intelligent large model system and control method thereof
By employing complex differential privacy algorithms and adversarial training mechanisms, combined with generative adversarial networks, the contradiction between privacy protection and data accuracy in medical aesthetics AI systems has been resolved. This has enabled the efficient generation and security protection of personalized medical aesthetics solutions, thereby improving the safety and user experience of medical aesthetics services.
Patent Information
- Application Number
- CN202411335443.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-24
- Publication Date
- 2025-01-07
AI Technical Summary
When existing differential privacy technology is used to add noise to protect privacy in medical aesthetics AI systems, it may lead to a decrease in data accuracy, making it difficult to balance the contradiction between privacy protection and data accuracy, and there are also security risks of malicious attacks and data tampering.
Employing a complex differential privacy algorithm and adversarial training mechanism, this system dynamically adjusts noise amplitude and implements multi-layered data security protection. Combined with generative adversarial networks, it generates personalized medical aesthetic solutions, monitors and defends against malicious attacks in real time, and optimizes the parameters of the generated model.
While protecting privacy, it ensures the accuracy and safety of cosmetic medical solutions, improves the efficiency and personalization of cosmetic medical services, and enhances the robustness and data security of the system.
Smart Images
Figure CN119272330A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a medical beauty AIGC intelligent large model system and a control method thereof. Background Art
[0002] In recent years, with the rapid development of artificial intelligence technology and generative adversarial networks, the medical beauty industry has gradually introduced AI technology into the medical beauty service process to improve efficiency, personalized service capabilities and customer satisfaction. AIGC technology can automatically generate customized medical beauty solutions based on the personalized needs of users through generative adversarial networks, thereby greatly improving the quality and response speed of medical beauty services.
[0003] When generating personalized plans, medical aesthetic systems often rely on a large amount of medical record data, which is extremely sensitive and directly affects the correctness of medical decisions. Malware intrusion or algorithm design defects may cause data tampering and even threaten the health and safety of patients. Especially when AI systems encounter adversarial attacks, attackers can mislead AI models by injecting defective data, ultimately making wrong medical aesthetic decisions. In addition, with the increasing awareness of data privacy protection, how to effectively protect patient privacy in the process of generating medical aesthetic plans has become a key issue in the development of AI medical aesthetic systems.
[0004] Differential privacy technology has been widely used in the field of data protection. It can ensure the privacy of data during the processing process by adding noise. However, how to maintain data accuracy in the process of adding noise and balance the contradiction between privacy protection and data accuracy are the difficulties currently faced by differential privacy technology. Traditional differential privacy algorithms may introduce too much noise, resulting in a decrease in the accuracy of medical beauty solutions. Therefore, a medical beauty AIGC intelligent large model system and its control method are proposed to address the above problems. Summary of the invention
[0005] The purpose of the present invention is to provide a medical beauty AIGC intelligent large model system and its control method to solve the problem that differential privacy technology has been widely used in the field of data protection. The privacy of data during the processing process can be ensured by adding noise. However, how to maintain data accuracy in the process of adding noise and balance the contradiction between privacy protection and data accuracy are the difficulties currently faced by differential privacy technology. Traditional differential privacy algorithms may introduce too much noise, resulting in a decrease in the accuracy of medical beauty solutions.
[0006] To achieve the above object, the present invention provides the following technical solutions:
[0007] A medical beauty AIGC intelligent large model system and a control method thereof, comprising a data acquisition module, a preprocessing and differential privacy module, an AIGC generation module, a data security and protection module and a control module. The data acquisition module collects the patient's medical data and beauty demand information through sensors and cameras;
[0008] The preprocessing and differential privacy module is used to clean and standardize the raw data, and to add noise using differential privacy technology to protect user privacy. The algorithm is used to dynamically adjust the amount of noise to find a balance between privacy protection and model performance.
[0009] The AIGC generation module generates a personalized medical aesthetic design plan based on a generative AI model of deep learning;
[0010] The data security and protection module uses encryption, malware detection, adversarial attack defense and data integrity verification mechanisms;
[0011] The control module controls the medical beauty device to perform operations according to the generated medical beauty plan, and adjusts the operating parameters through a closed-loop feedback mechanism;
[0012] The system achieves data privacy protection and security protection through differential privacy technology and multi-layer data security mechanism, and generates personalized medical beauty solutions based on the AIGC model.
[0013] As a further optimization of the present invention, the preprocessing and differential privacy module adopts a complicated differential privacy algorithm when generating data noise, and the formula is:
[0014]
[0015] Where M(D) is the data result after adding noise, is the Gaussian noise of the data point, φ(i,D) is the weighting function based on the data features, and the integral term introduces dynamic noise adjustment to enhance the data privacy protection effect.
[0016] As a further optimization of the present invention, the differential privacy module dynamically adjusts the noise amplitude according to the data sensitivity, using the following formula:
[0017]
[0018] Where △f is the data sensitivity and ε is the privacy budget.
[0019] As a further optimization of the present invention, the AIGC generation module uses a generative adversarial network (GAN) to generate medical beauty solutions, generates personalized medical beauty solutions through the collaborative work of the generator and the discriminator, and evaluates the quality of the generated solutions through the discriminator to ensure that the solutions meet the needs of users.
[0020] As further optimized content of the present invention, the AIGC generation module can adjust the parameters of the generation model according to real-time feedback, and further improve the accuracy of medical beauty solution generation through an optimization algorithm based on reinforcement learning.
[0021] As a further optimized content of the present invention, the data security and protection module includes a security protection mechanism based on adversarial training to prevent malware intrusion, data tampering and adversarial attacks.
[0022] As further optimized content of the present invention, the adversarial protection mechanism enhances the robustness of the system in the face of malicious attacks through adversarial training, and can detect and defend against adversarial attacks or data tampering in real time.
[0023] As a further optimization of the present invention, the control module can accurately control the medical beauty equipment based on the medical beauty plan generated by AIGC, and adjust the operating parameters in real time through the feedback mechanism.
[0024] As a further optimized content of the present invention, the following steps are included:
[0025] Step 1: Collect multi-dimensional data such as emotional state, body temperature, physical activity and blood oxygen level: Through wearable devices and facial expression recognition, voice emotion analysis, and heart rate change sensors, collect the user's emotional state data in real time to reflect the user's mental and emotional fluctuations. Use infrared temperature sensors to obtain the user's body surface temperature to evaluate the user's physiological state. Use smart bracelets to record the user's daily physical activity and exercise patterns, obtain the user's steps and heart rate, and use pulse oximeters to collect the user's blood oxygen concentration. Combined with other physiological indicators, provide health information reference for the generation of medical beauty plans;
[0026] Step 2: Preprocess the collected data and use a complex differential privacy algorithm: clean the collected raw data to remove noise, abnormal data or incomplete data, and then standardize the data with different characteristics to ensure that the data in each dimension has a consistent scale and distribution;
[0027] Noise is added to user data through the differential privacy algorithm to ensure data privacy security. The formula is:
[0028]
[0029] In the formula, Represents Gaussian noise, combining a weighted function of each data point’s features and an additional integral term to achieve a complex noise distribution, further enhancing privacy protection while retaining the actual validity of the data;
[0030] Step 3: Generate personalized medical beauty solutions based on the AIGC generation module: Based on the collected data such as user emotions, body temperature, and physical activity, the system inputs this information into the AIGC generation module and uses the generative adversarial network to generate personalized medical beauty solutions. The generator generates solutions based on the input data, while the discriminator evaluates the quality of the generated solutions to ensure that the generated medical beauty solutions meet the personalized needs of users and have sufficient accuracy.
[0031] Step 4: Detect and respond to potential malicious attacks through adversarial protection mechanisms: The system uses the protection module to perform real-time data security monitoring to check for abnormal data input, malware intrusion, or potential adversarial attacks. The monitoring mechanism uses adversarial protection technology to identify and resist attackers' attempts to tamper with or mislead the AI system.
[0032] The protection module enhances the robustness of the system under malicious attacks through adversarial training. Adversarial training introduces noise or hostile samples into the model for training, so that the model can identify potential malicious inputs and maintain the stability of medical decision-making when attacked.
[0033] Step 5: Control the medical beauty equipment to perform operations according to the plan, and adjust parameters through real-time feedback: Based on the generated personalized medical beauty plan, the system instructs the relevant medical beauty equipment to perform specific operation steps through the control module. These operations are performed automatically by the equipment. During the operation, the sensors installed on the equipment will monitor the user's skin condition and physiological indicators in real time, and feed this data back to the control module. The control module dynamically adjusts the operating parameters of the medical beauty equipment based on the feedback data. Through the data fed back by the sensors, the system optimizes the various parameters of energy intensity and operation time during the execution of the medical beauty plan.
[0034] As a further optimization content of the present invention, the solution generated in step 3 is continuously optimized by adjusting the parameters of the generator through multiple cycles of iterations.
[0035] Compared with the prior art, the present invention has the following beneficial effects:
[0036] 1. In the present invention, the differential privacy technology is combined with a multi-layer protection mechanism of adversarial training to effectively solve the potential security risks and privacy protection problems that the medical beauty AI system may face when processing medical record data. Through the improved differential privacy algorithm, the system can dynamically adjust the noise amplitude according to the sensitivity of the data, and the formula is further complicated to ensure that the accuracy of the data is retained to the greatest extent while adding noise. This enables the system to generate accurate medical beauty solutions while protecting privacy. Through the integration of advanced technologies such as privacy protection, adversarial training, and generative adversarial networks, it not only significantly improves the efficiency and personalization level of medical beauty services, but also provides strong security protection for medical beauty data, effectively addressing the privacy leakage and security issues of AI technology in medical data processing;
[0037] 2. In the present invention, through adversarial training, the system can effectively detect and resist adversarial attacks and malware intrusions, prevent attackers from misleading the AI model by injecting defective data, thereby ensuring the accuracy and security of the generated medical beauty solutions. At the same time, based on the generative adversarial network of the AIGC model, the system can generate personalized medical beauty solutions based on multi-dimensional patient data. Through the reinforcement learning algorithm, the system can also adjust the parameters of the generated model in real time according to user feedback, continuously optimize the generated solutions, and improve the accuracy and personalized experience of medical beauty services;
[0038] 3. In the present invention, the system has built-in multi-layer data security protection modules, including malware protection, data tampering detection and adversarial protection mechanisms, which greatly improve the robustness of the system in the face of complex network attacks. Through complex adversarial protection algorithms, the system can maintain normal operation in the face of malicious attacks, ensuring the safety and accuracy of the medical beauty process. The system can fully acquire the user's physiological and psychological data through a multi-dimensional data acquisition module, and standardize and clean it through a data preprocessing module to ensure the reliability of subsequent generation plans. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] Figure 1 This is a flow chart of a medical beauty AIGC intelligent large model system and its control method of the present invention. DETAILED DESCRIPTION
[0040] See also Figure 1 , the present invention provides a technical solution:
[0041] A medical beauty AIGC intelligent large model system and a control method thereof, comprising a data acquisition module, a preprocessing and differential privacy module, an AIGC generation module, a data security and protection module and a control module. The data acquisition module collects the patient's medical data and beauty demand information through sensors and cameras; the preprocessing and differential privacy module is used to clean and standardize the original data, and apply differential privacy technology to add noise to protect user privacy, and adopt an algorithm to dynamically adjust the amount of noise to find a balance between privacy protection and model performance; the AIGC generation module generates personalized medical beauty design solutions based on a generative AI model of deep learning; the data security and protection module uses encryption, malware detection, adversarial attack defense and data integrity verification mechanisms; the control module controls the medical beauty equipment to perform operations according to the generated medical beauty solutions, and adjusts the operating parameters through a closed-loop feedback mechanism; the system realizes data privacy protection and security protection through differential privacy technology and multi-layer data security mechanism, and generates personalized medical beauty solutions based on the AIGC model. This system ensures the security and privacy of user data through differential privacy technology, and combines AIGC to generate personalized medical beauty solutions, which greatly improves the personalization and efficiency of services and meets the personalized needs of users;
[0042] As a technical solution for further implementation of this solution, the preprocessing and differential privacy module uses a complex differential privacy algorithm when generating data noise. The formula is:
[0043]
[0044] Where M(D) is the data result after adding noise, is the Gaussian noise of the data point, φ(i,D) is a weighted function based on data features, and the integral term introduces dynamic noise adjustment to enhance the data privacy protection effect. The complex algorithm enhances the security of the data and further ensures that the privacy data is not leaked during the processing process, especially when generating medical beauty plans, it can also maintain high accuracy and personalization;
[0045] As a technical solution for further implementation of this solution, the differential privacy module dynamically adjusts the noise amplitude according to the data sensitivity, using the following formula:
[0046]
[0047] In the formula, △f is the data sensitivity and ε is the privacy budget. This dynamic noise adjustment mechanism enables the system to flexibly adapt to the sensitivity of different data, ensuring that sensitive data can still maintain high-precision output under privacy protection and avoiding unnecessary privacy leakage risks.
[0048] As a technical solution for further implementation of this solution, the AIGC generation module uses a generative adversarial network (GAN) to generate medical beauty solutions. It generates personalized medical beauty solutions through the collaborative work of the generator and the discriminator, and evaluates the quality of the generated solutions through the discriminator to ensure that the solutions meet the needs of users. The GAN model can generate more realistic and personalized medical beauty solutions through the collaborative work of the generator and the discriminator, improve the user experience, and ensure the feasibility and high quality of the solutions.
[0049] As a technical solution for further implementation of this solution, the AIGC generation module can adjust the parameters of the generation model according to real-time feedback, and further improve the accuracy of medical beauty solution generation through an optimization algorithm based on reinforcement learning. Through the introduction of real-time feedback and reinforcement learning, the system can continuously optimize the accuracy and personalization of medical beauty solution generation;
[0050] As a technical solution for further implementation of this plan, the data security and protection module includes a security protection mechanism based on adversarial training to prevent malware intrusion, data tampering and adversarial attacks, effectively resist various malicious attacks and intrusions, ensure the security and integrity of data, especially in sensitive medical environments, greatly improving the reliability and trust of the system;
[0051] As a technical solution for further implementation of this plan, the adversarial protection mechanism enhances the robustness of the system in the face of malicious attacks through adversarial training, and can detect and defend against adversarial attacks or data tampering in real time. The adversarial protection mechanism enhances the system's anti-attack capability, ensures that data is not tampered with or misled under any circumstances, improves the robustness of the system, and ensures the security of user data and medical beauty solutions;
[0052] As a technical solution for further implementation of this solution, the control module can accurately control the medical beauty equipment based on the medical beauty solution generated by AIGC, and adjust the operating parameters in real time through the feedback mechanism;
[0053] As a technical solution for further implementing this plan, the following steps are included:
[0054] Step 1: Collect multi-dimensional data such as emotional state, body temperature, physical activity and blood oxygen level: Through wearable devices and facial expression recognition, voice emotion analysis, and heart rate change sensors, collect the user's emotional state data in real time to reflect the user's mental and emotional fluctuations. Use infrared temperature sensors to obtain the user's body surface temperature to evaluate the user's physiological state. Use smart bracelets to record the user's daily physical activity and exercise patterns, obtain the user's steps and heart rate, and use pulse oximeters to collect the user's blood oxygen concentration. Combined with other physiological indicators, provide health information reference for the generation of medical beauty plans;
[0055] Step 2: Preprocess the collected data and use a complex differential privacy algorithm: clean the collected raw data to remove noise, abnormal data or incomplete data, and then standardize the data with different characteristics to ensure that the data in each dimension has a consistent scale and distribution;
[0056] Noise is added to user data through the differential privacy algorithm to ensure data privacy security. The formula is:
[0057]
[0058] In the formula, Represents Gaussian noise, combining a weighted function of each data point’s features and an additional integral term to achieve a complex noise distribution, further enhancing privacy protection while retaining the actual validity of the data;
[0059] Step 3: Generate personalized medical beauty solutions based on the AIGC generation module: Based on the collected data such as user emotions, body temperature, and physical activity, the system inputs this information into the AIGC generation module and uses the generative adversarial network to generate personalized medical beauty solutions. The generator generates solutions based on the input data, while the discriminator evaluates the quality of the generated solutions to ensure that the generated medical beauty solutions meet the personalized needs of users and have sufficient accuracy.
[0060] Step 4: Detect and respond to potential malicious attacks through adversarial protection mechanisms: The system uses the protection module to perform real-time data security monitoring to check for abnormal data input, malware intrusion, or potential adversarial attacks. The monitoring mechanism uses adversarial protection technology to identify and resist attackers' attempts to tamper with or mislead the AI system.
[0061] The protection module enhances the robustness of the system under malicious attacks through adversarial training. Adversarial training introduces noise or hostile samples into the model for training, so that the model can identify potential malicious inputs and maintain the stability of medical decision-making when attacked.
[0062] Step 5: Control the medical beauty equipment to perform operations according to the plan, and adjust parameters through real-time feedback: Based on the generated personalized medical beauty plan, the system instructs the relevant medical beauty equipment to perform specific operation steps through the control module. These operations are performed automatically by the equipment. During the operation, the sensors installed on the equipment will monitor the user's skin condition and physiological indicators in real time, and feed back these data to the control module. The control module dynamically adjusts the operating parameters of the medical beauty equipment according to the feedback data. Through the data fed back by the sensors, the system optimizes the energy intensity and operation time parameters during the execution of the medical beauty plan. The plan generated in step 3 adjusts the parameters of the generator through multiple cycles of iteration to continuously optimize the generated plan. This method ensures the security and accuracy of the entire process from data collection to plan execution. Through multiple security protections such as differential privacy and adversarial training, it ensures the effectiveness and safety of the medical beauty plan, improves user experience and ensures data privacy.
[0063] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the method and core ideas of the present invention. The above are only preferred implementation methods of the present invention. It should be pointed out that due to the limitations of textual expression, there are objectively infinite specific structures. For ordinary technicians in this technical field, without departing from the principles of the present invention, they can make several improvements, modifications or changes, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes or combinations, or the direct application of the inventive concept and technical solution to other occasions without improvement, should be regarded as the protection scope of the present invention.
Claims
1. A medical beauty AIGC intelligent large model system, including a data acquisition module, a preprocessing and differential privacy module, an AIGC generation module, a data security and protection module and a control module, characterized in that: The data acquisition module collects the patient's medical data and beauty demand information through sensors and cameras; The preprocessing and differential privacy module is used to clean and standardize the raw data, and to add noise using differential privacy technology to protect user privacy. The algorithm is used to dynamically adjust the amount of noise to find a balance between privacy protection and model performance. The AIGC generation module generates a personalized medical aesthetic design plan based on a generative AI model of deep learning; The data security and protection module uses encryption, malware detection, adversarial attack defense and data integrity verification mechanisms; The control module controls the medical beauty device to perform operations according to the generated medical beauty plan, and adjusts the operating parameters through a closed-loop feedback mechanism; The system achieves data privacy protection and security protection through differential privacy technology and multi-layer data security mechanism, and generates personalized medical beauty solutions based on the AIGC model.
2. According to claim 1, a medical beauty AIGC intelligent large model system is characterized by: The preprocessing and differential privacy module uses a complex differential privacy algorithm when generating data noise. The formula is: Where M(D) is the data result after adding noise, is the Gaussian noise of the data point, φ(i,D) is the weighting function based on the data features, and the integral term introduces dynamic noise adjustment to enhance the data privacy protection effect.
3. According to claim 1, the medical beauty AIGC intelligent large model system is characterized by: The differential privacy module dynamically adjusts the noise amplitude according to the data sensitivity, using the following formula: Where △f is the data sensitivity and ε is the privacy budget.
4. According to claim 1, the medical beauty AIGC intelligent large model system is characterized by: The AIGC generation module uses a generative adversarial network (GAN) to generate medical beauty solutions. It generates personalized medical beauty solutions through the collaborative work of the generator and the discriminator, and evaluates the quality of the generated solutions through the discriminator to ensure that the solutions meet the needs of users.
5. According to claim 1, the medical beauty AIGC intelligent large model system is characterized by: The AIGC generation module can adjust the parameters of the generation model according to real-time feedback, and further improve the accuracy of medical beauty plan generation through an optimization algorithm based on reinforcement learning.
6. According to claim 1, the medical beauty AIGC intelligent large model system is characterized by: The data security and protection module includes a security protection mechanism based on adversarial training to prevent malware intrusion, data tampering and adversarial attacks.
7. According to claim 1, the medical beauty AIGC intelligent large model system is characterized by: The adversarial protection mechanism enhances the robustness of the system in the face of malicious attacks through adversarial training, and can detect and defend against adversarial attacks or data tampering in real time.
8. According to claim 1, the medical beauty AIGC intelligent large model system is characterized by: The control module can accurately control the medical beauty equipment based on the medical beauty plan generated by AIGC, and adjust the operating parameters in real time through the feedback mechanism.
9. A control method for a medical beauty AIGC intelligent large model system according to any one of claims 1 to 8, characterized in that: The following steps are involved: Step 1: Collect multi-dimensional data such as emotional state, body temperature, physical activity and blood oxygen level: Through wearable devices and facial expression recognition, voice emotion analysis, and heart rate change sensors, collect the user's emotional state data in real time to reflect the user's mental and emotional fluctuations. Use infrared temperature sensors to obtain the user's body surface temperature to evaluate the user's physiological state. Use smart bracelets to record the user's daily physical activity and exercise patterns, obtain the user's steps and heart rate, and use pulse oximeters to collect the user's blood oxygen concentration. Combined with other physiological indicators, provide health information reference for the generation of medical beauty plans; Step 2: Preprocess the collected data and use a complex differential privacy algorithm: clean the collected raw data to remove noise, abnormal data or incomplete data, and then standardize the data with different characteristics to ensure that the data in each dimension has a consistent scale and distribution; Noise is added to user data through the differential privacy algorithm to ensure data privacy security. The formula is: In the formula, Represents Gaussian noise, combining a weighted function of each data point’s features and an additional integral term to achieve a complex noise distribution, further enhancing privacy protection while retaining the actual validity of the data; Step 3: Generate personalized medical beauty solutions based on the AIGC generation module: Based on the collected data such as user emotions, body temperature, and physical activity, the system inputs this information into the AIGC generation module and uses the generative adversarial network to generate personalized medical beauty solutions. The generator generates solutions based on the input data, while the discriminator evaluates the quality of the generated solutions to ensure that the generated medical beauty solutions meet the personalized needs of users and have sufficient accuracy. Step 4: Detect and respond to potential malicious attacks through adversarial protection mechanisms: The system uses the protection module to perform real-time data security monitoring to check for abnormal data input, malware intrusion, or potential adversarial attacks. The monitoring mechanism uses adversarial protection technology to identify and resist attackers' attempts to tamper with or mislead the AI system. The protection module enhances the robustness of the system under malicious attacks through adversarial training. Adversarial training introduces noise or hostile samples into the model for training, so that the model can identify potential malicious inputs and maintain the stability of medical decision-making when attacked. Step 5: Control the medical beauty equipment to perform operations according to the plan, and adjust parameters through real-time feedback: Based on the generated personalized medical beauty plan, the system instructs the relevant medical beauty equipment to perform specific operation steps through the control module. These operations are performed automatically by the equipment. During the operation, the sensors installed on the equipment will monitor the user's skin condition and physiological indicators in real time, and feed this data back to the control module. The control module dynamically adjusts the operating parameters of the medical beauty equipment based on the feedback data. Through the data fed back by the sensors, the system optimizes the various parameters of energy intensity and operation time during the execution of the medical beauty plan.
10. The control method of the medical beauty AIGC intelligent large model system according to claim 9 is characterized by: The solution generated in step 3 is continuously optimized by adjusting the parameters of the generator through multiple cycles of iterations.