Method and system for evaluating facial skin aging

Abstract

The invention provides a method and system for evaluating facial skin aging, and the method comprises the steps: obtaining visible light, ultraviolet reflection and near-infrared spectrum images through multispectral imaging, and extracting skin texture, color spots and capillary network parameters through preprocessing; constructing a three-dimensional facial skin model, and measuring skin elasticity parameters in combination with a mechanical sensor; and inputting the parameters into a deep learning aging evaluation model, outputting a comprehensive evaluation report containing the physiological age of the skin, the elasticity grade and the regional aging difference index, and providing personalized aging protection suggestions. According to the method, multi-modal data and a deep learning technology are fused, multi-dimensional quantitative evaluation is realized, the method has the characteristics of high precision, objectivity and individuation, a powerful tool is provided for skin health management, and the method has a wide application prospect.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of dermatology and biomedical engineering, and particularly relates to a method and system for evaluating facial skin aging. BACKGROUND

[0002] Skin aging is a complex physiological process influenced by various internal and external factors, including natural aging, ultraviolet radiation, environmental pollution, lifestyle, etc. Traditional methods of skin aging assessment rely heavily on subjective visual evaluation or single physiological indicator detection, such as skin elasticity testing, wrinkle depth measurement, etc. These methods have the disadvantages of strong subjectivity, limited evaluation dimensions, and difficulty in fully reflecting the skin aging condition. In recent years, with the rapid development of imaging technology, sensor technology and artificial intelligence algorithms, it has provided new possibilities for multi-dimensional, objective and quantitative evaluation of skin aging. SUMMARY

[0003] The present application proposes a comprehensive method for evaluating facial skin aging, which realizes multi-dimensional quantitative evaluation of skin aging through multi-spectral imaging, mechanical sensing, deep learning and other technical means, and generates a comprehensive evaluation report containing skin physiological age, elasticity grade and regional aging difference index, while providing personalized aging prevention suggestions, providing a powerful tool for skin health management.

[0004] I. Method steps: S1 Multi-dimensional image data acquisition A multi-spectral imaging device is used, including a visible light camera, an ultraviolet reflection camera and a near-infrared spectrum camera, to acquire visible light images, ultraviolet reflection images and near-infrared spectrum images of facial skin respectively. Visible light images are used to observe the overall appearance of the skin, including wrinkles, pigmented spots and other macroscopic features; ultraviolet reflection images can reveal the fine structure and pigment distribution of the skin surface; near-infrared spectrum images can provide spectral information of the deep tissue of the skin, reflecting changes in water content, collagen and other components of the skin. Through multi-spectral imaging, rich information of the skin can be obtained from different wavebands and different levels, laying a foundation for subsequent feature extraction and analysis.

[0005] S2 Image data preprocessing and feature parameter extraction The obtained multi-dimensional image data is pre-processed, mainly including standardization registration and noise elimination processing, to ensure the consistency and accuracy of the image data. The watershed algorithm is used to segment the skin spot area, which can effectively identify and separate the spot area, providing the basis for the extraction of spot distribution parameters. At the same time, the histogram of oriented gradients (HOG) is used to extract wrinkle distribution features. The HOG algorithm can capture the direction and intensity information of the wrinkle, so as to quantitatively describe the distribution of the wrinkle. In addition, capillary network parameters need to be extracted, and image analysis technology is used to identify the distribution, density and morphology of capillaries, etc. The changes of capillary network are closely related to skin aging, such as capillary dilation and density reduction, which are often the early manifestations of skin aging.

[0006] S3 Three-dimensional facial skin model construction and elasticity parameter acquisition A three-dimensional facial skin model is constructed using a three-dimensional scanner, which can accurately reflect the geometric shape and surface texture of the facial skin. A vertical pressure of 0.5-1.2N is applied to the pre-set test points on the face, and the deformation process of the skin during the pressure process is recorded by a high frame rate camera (≥1000fps). Based on the viscoelastic model, the recorded deformation data is analyzed to calculate the stress relaxation time constant τ and the elastic recovery rate η, which are important indicators for measuring skin elasticity and can reflect the recovery ability and elastic retention ability of the skin after being subjected to external force. Skin elasticity gradually decreases with age, which is one of the significant characteristics of skin aging. By accurately measuring the elasticity parameters, key mechanical basis can be provided for skin aging assessment.

[0007] S4 Application of deep learning aging assessment model The skin texture feature parameters, the color spot distribution parameters, the capillary network parameters and the elasticity parameters obtained in steps S2 and S3 are input into an aging assessment model trained by deep learning. The model adopts a multi-task learning architecture and contains three branch networks: the first branch network is dedicated to processing texture features and color spot parameters, extracts multi-scale texture and color spot features through a convolutional neural network (CNN), and compares them with an age-related database; the second branch network processes capillary network topology features, analyzes the distribution pattern and morphological changes of capillary blood vessels, and further excavates the blood vessel features related to aging; the third branch network processes elasticity mechanics parameters, converts parameters such as stress relaxation time and elastic recovery rate into quantitative indicators related to the degree of aging. Through the attention mechanism, the features output by each branch are fused, so that the model can comprehensively consider the influence of different features on skin aging, improve the accuracy and comprehensiveness of the assessment. In the model training stage, a transfer learning method is adopted, and a multi-modal skin database containing 10,000 cases of different races, genders and ages is used for pre-training, so as to fully utilize the existing large-scale data resources, accelerate the training process of the model and improve the generalization ability of the model. In addition, an adversarial generative network (GAN) is used to enhance the training data, generate synthetic data under different lighting conditions and skin states, further enrich the diversity of training samples, enhance the adaptability of the model to various complex situations, and improve the robustness of the model in actual application.

[0008] S5 Comprehensive assessment report output and personalized recommendation generation The comprehensive assessment report containing the skin physiological age, the elasticity grade and the regional aging difference index is output by comprehensively considering the parameters obtained in steps S2 and S3 and the assessment results of the deep learning model. The skin physiological age is an aging degree index corresponding to the actual age, which is obtained by comprehensively evaluating various features of the skin, and can more directly reflect the actual aging state of the skin; the elasticity grade is classified according to the skin elasticity parameters, and provides specific evaluation of the skin elasticity for the user; the regional aging difference index is obtained by the formula DI=(Emax-Emin) / Eavg x 100%, wherein Emax is the maximum aging value in each subarea of the face, Emin is the minimum aging value, and Eavg is the average aging value. This index can reveal the difference in the aging degree of different regions of the face, helping the user to understand the distribution of skin aging. In addition, individual parameters input by the user, including age, gender, geographical residence information and skin care habits, are input into the assessment model together with the detection parameters to generate personalized aging prevention suggestions and provide targeted skin care guidance for the user, such as recommending suitable skin care products, adjusting skin care habits, and providing lifestyle improvement suggestions, so as to delay the skin aging process.

[0009] II. System composition The application also provides an evaluation system for implementing the above method, comprising the following modules: Image acquisition module: contains a multi-spectral imaging device and a three-dimensional scanner, used to obtain multi-dimensional image data and three-dimensional model data of facial skin.

[0010] Data processing module: configured with GPU accelerated computing unit, can efficiently process a large amount of image data and conduct complex deep learning calculation, ensure the fast response and accurate evaluation of the system.

[0011] Parameter database: stores skin feature reference values based on age grouping, provides standard data support for evaluation model, facilitates comparative analysis of test results, and determines the degree of skin aging.

[0012] Result output module: supports visual charts and augmented reality (AR) simulation display, can show the evaluation results to the user in an intuitive way, such as displaying the aging degree distribution map, the comparison between skin physiological age and actual age on the user's face in real time through AR technology, etc., to enhance user experience and understanding of the evaluation results. Advantages

[0013] The advantages of the present application mainly include the following aspects: (1) Multi-dimensional evaluation: by integrating multi-spectral imaging, mechanical sensing and deep learning technologies, comprehensive evaluation is carried out from multiple dimensions such as skin texture, spots, capillary network and elasticity, which fully reflects the skin aging condition and overcomes the limitations of single index evaluation in traditional methods.

[0014] (2) High precision and objectivity: advanced imaging technology and deep learning algorithm are used to accurately extract skin features and conduct quantitative analysis, reducing human factor interference and improving the objectivity and accuracy of evaluation.

[0015] (3) Personalized recommendations: personalized aging prevention recommendations are generated based on individual user parameters, providing targeted skin care guidance to help users take effective anti-aging measures according to their own conditions.

[0016] (4) Dynamic prediction: dynamic aging prediction curve is generated in the evaluation report, based on the current parameters to establish a skin state differential equation, which can predict the future aging trend of the skin and provide a reference for users to plan skin care programs in advance.

[0017] (5) System integration and convenience: the supporting evaluation system integrates image acquisition, data processing, result output and other functional modules, easy to operate and easy to popularize and apply in medical institutions, beauty institutions and other places, providing a one-stop solution for skin health management. DETAILED DESCRIPTION

[0018] Example 1: 35-year-old female subject Image data acquisition: The face of a 35-year-old female subject was photographed using a multispectral imaging device, obtaining visible light images, ultraviolet reflection images, and near-infrared spectral images respectively. The visible light image shows that the subject's face has a small number of fine lines and slight pigmentation; the ultraviolet reflection image reveals the fine structure of the pigmented area, with a clear edge; the near-infrared spectral image reflects the spectral information of the deep tissue of the skin, and preliminary judgment shows that the skin moisture content is moderate, but there are signs of slight loss of collagen components.

[0019] Image data preprocessing and feature parameter extraction: The acquired multi-dimensional image data is standardized and registered, and noise is removed. The watershed algorithm is used to segment the pigmented area, and the pigmented area distribution parameters are calculated, such as the pigmented area ratio of 2.5% and the pigmented density of 3.2 per square centimeter. The wrinkle distribution characteristics are extracted by the histogram of oriented gradients (HOG), and the wrinkle distribution intensity is 0.45 (normalized value), mainly concentrated in the eye and mouth area. At the same time, the capillary network parameters are extracted, and it is found that the capillary density is 50 per square millimeter, slightly lower than that of younger skin, and some capillaries appear slightly dilated.

[0020] Three-dimensional facial skin model construction and elasticity parameter acquisition: A three-dimensional scanner is used to construct the three-dimensional facial skin model of the subject, and a vertical pressure of 0.8N is applied to multiple preset test points on the face. The skin deformation process is recorded by a high-frame-rate camera. Based on the viscoelastic model, the stress relaxation time constant τ of the subject's skin is calculated to be 1.2 seconds, and the elastic recovery rate η is 75%, indicating that the skin elasticity has decreased compared to younger skin, but is still within the normal range.

[0021] Deep learning aging assessment model application: The extracted feature parameters are input into the deep learning trained aging assessment model, and the model analyzes the parameters through a multi-task learning architecture. The first branch network processes texture features and pigmented area parameters, and obtains a texture aging index associated with age of 0.6; the second branch network processes capillary network topology features, and obtains a vascular aging index of 0.4; the third branch network processes elastic mechanics parameters, and obtains an elastic aging index of 0.5. Through the attention mechanism, the output features of each branch are fused, and finally the subject's skin physiological age is calculated to be 38 years old, the elasticity level is medium, and the regional aging difference index is 15%.

[0022] Comprehensive evaluation report output and personalized recommendation generation: According to the evaluation results, a comprehensive evaluation report is output, showing that the subject's skin physiological age is slightly higher than the actual age, the elasticity is at a medium level, and the facial aging degree is more obvious in the eye and mouth corner areas. At the same time, the individual parameters input by the subject are received, including age 35 years old, female, living in the city, good daily skin care habits but lack of sun protection measures, etc. These individual parameters and detection parameters are jointly input into the evaluation model to generate personalized anti-aging recommendations, suggesting that the subject should strengthen sun protection measures, use skin care products containing antioxidant ingredients, and regularly perform eye and lip care to delay the skin aging process.

[0023] Example 2: 40-year-old male user System building: A complete evaluation system is built, including an image acquisition module, a data processing module, a parameter database, and a result output module. The image acquisition module is equipped with a multi-spectral imaging device and a three-dimensional scanner to ensure that high-quality image data can be obtained; the data processing module is configured with a high-performance GPU accelerated computing unit to meet the computing needs of the deep learning model; the parameter database pre-stores skin feature reference values based on age grouping, covering skin data of different age groups, genders, and races; the result output module supports visual charts and augmented reality (AR) simulation display to provide users with intuitive evaluation result display.

[0024] System operation process: A 40-year-old male user comes to a beauty institution for skin aging evaluation. First, the staff uses the image acquisition module in the system to perform multi-spectral imaging and three-dimensional scanning on the user's face to obtain facial image data and three-dimensional model data. Then, the system automatically preprocesses the image data and extracts feature parameters, and simultaneously obtains skin elasticity parameters through a mechanical sensor. The extracted parameters and elasticity data are input into the deep learning aging evaluation model, which outputs a comprehensive evaluation report after rapid calculation. The report shows that the user's skin physiological age is 45 years old, the elasticity level is poor, the regional aging difference index is 20%, and the facial aging degree is more significant in the forehead and neck. The user intuitively sees the distribution of facial aging and the gap with the actual age through the visual charts and AR simulation display of the result output module. Finally, the user adjusts his skin care habits according to the personalized anti-aging recommendations generated by the system, increases neck care and the use of anti-aging skin care products, and regularly goes to the beauty institution for follow-up evaluation.

[0025] Example 3: 50-year-old female subject Image data acquisition: Multispectral imaging was performed on a 50-year-old female subject to obtain visible light images, ultraviolet reflectance images, and near-infrared spectral images. The visible light images showed that the subject's face had obvious wrinkles and more color spots; the ultraviolet reflectance images revealed the fine structure of the color spot area, and the color spot edges were relatively blurred, indicating that the color spots were more serious; the near-infrared spectral images showed the spectral information of the deep tissue of the skin, indicating that the skin moisture content was low, and the collagen component loss was more obvious.

[0026] Image data preprocessing and feature parameter extraction: The image data was standardized and registered, and noise elimination processing was performed, and the color spot area was segmented using the watershed algorithm, and the color spot area ratio was calculated to be 5.0%, and the color spot density was 5.0 per square centimeter; the wrinkle distribution characteristics were extracted by the histogram of oriented gradients (HOG), and the wrinkle distribution intensity was obtained to be 0.75 (normalized value), and the wrinkles were mainly concentrated in the forehead, eye and mouth corner regions; At the same time, the capillary network parameters were extracted, and it was found that the capillary density was 40 per square millimeter, and the capillary expansion phenomenon was more obvious.

[0027] Three-dimensional facial skin model construction and elasticity parameter acquisition: A three-dimensional facial skin model of the subject was constructed using a three-dimensional scanner, and a vertical pressure of 1.0 N was applied to multiple preset test points on the face, and the skin deformation process was recorded by a high-frame-rate camera. Based on the viscoelastic model, it was calculated that the stress relaxation time constant τ of the subject's skin was 1.5 seconds, and the elastic recovery rate η was 60%, indicating that the skin elasticity decreased significantly.

[0028] Deep learning aging assessment model application: The extracted feature parameters are input into the deep learning aging assessment model, and the model analyzes each parameter through a multi-task learning architecture. The first branch network processes texture features and color spot parameters, and obtains a texture aging index associated with age of 0.8; the second branch network processes capillary network topology features, and obtains a vascular aging index of 0.6; the third branch network processes elastic mechanics parameters, and obtains an elastic aging index of 0.7. Through the attention mechanism, the output features of each branch are fused, and finally the subject's skin physiological age is obtained to be 55 years old, the elasticity level is poor, and the regional aging difference index is 25%.

[0029] Comprehensive evaluation report output and personalized recommendation generation: According to the evaluation results, output a comprehensive evaluation report showing that the subject's skin physiological age is significantly higher than the actual age, the elasticity is at a poor level, and the facial aging degree is more serious in the forehead, eye, and mouth corner areas. At the same time, receive the subject's input individual parameters, including age 50, female, living in the city, general daily skincare habits, and lack of exercise, etc. Input these individual parameters and detection parameters into the evaluation model together to generate personalized anti-aging recommendations, suggesting that the subject increase exercise, use skincare products with high moisturizing and anti-wrinkle efficacy, and regularly undergo professional skin care such as micro-needle treatment and photorejuvenation to improve skin condition and delay aging process.

[0030] Example 4: 60-year-old male subject Image data acquisition: Multi-spectral imaging was performed on a 60-year-old male subject to obtain visible light images, ultraviolet reflection images, and near-infrared spectral images. The visible light images showed that the subject had obvious and deep wrinkles on the face, and many and widely distributed color spots; the ultraviolet reflection images revealed the fine structure of the color spot area, with blurred color spot edges and obvious color spot fusion; the near-infrared spectral images showed the spectral information of the deep tissue of the skin, indicating that the skin had very low moisture content and severe loss of collagen composition.

[0031] Image data preprocessing and feature parameter extraction: The image data was standardized and registered, and noise was removed. The color spot area was segmented using the watershed algorithm, and the color spot area ratio was calculated to be 7.0%, and the color spot density was 6.0 per square centimeter. The wrinkle distribution characteristics were extracted by the histogram of oriented gradients (HOG), and the wrinkle distribution intensity was 0.85 (normalized value), with wrinkles mainly concentrated in the forehead, eye, nasolabial fold, and neck regions. The capillary network parameters were also extracted, and it was found that the capillary density was 30 per square millimeter, the capillary dilation was severe, and the capillary was broken in some areas.

[0032] Three-dimensional facial skin model construction and elasticity parameter acquisition: A three-dimensional facial skin model of the subject was constructed using a three-dimensional scanner, and a vertical pressure of 1.2 N was applied to multiple preset test points on the face. The skin deformation process was recorded by a high-frame-rate camera. Based on the viscoelastic model, the stress relaxation time constant τ of the subject's skin was calculated to be 1.8 seconds, and the elastic recovery rate η was 50%, indicating that the skin elasticity was significantly decreased and the elastic recovery ability was extremely poor.

[0033] Deep learning aging assessment model application: The extracted feature parameters are input into the deep learning aging assessment model, which comprehensively analyzes each parameter through a multi-task learning architecture. The first branch network processes texture features and speckle parameters, resulting in a texture aging index associated with age of 0.9; the second branch network processes capillary network topology features, resulting in a vascular aging index of 0.8; the third branch network processes elastomechanics parameters, resulting in an elasticity aging index of 0.8. Through the attention mechanism, the output features of each branch are fused, and finally the subject's skin physiological age is 68 years old, the elasticity level is extremely poor, and the regional aging difference index is 30%.

[0034] Comprehensive assessment report output and personalized recommendation generation: According to the assessment results, a comprehensive assessment report is output, showing that the subject's skin physiological age is much higher than the actual age, the elasticity is at an extremely poor level, and the facial aging degree is particularly serious in the forehead, eye area, nasolabial fold, and neck area. At the same time, the subject's input individual parameters are received, including age 60, male, living in rural areas, poor daily skincare habits, and long-term exposure to sunlight, etc. These individual parameters are input into the assessment model together with the detection parameters to generate personalized anti-aging recommendations, suggesting that the subject should strengthen sun protection measures, use skincare products with high moisturizing and repair efficacy, and regularly undergo professional skin care such as radiofrequency skin tightening and laser treatment to improve skin condition and slow down the aging process. In addition, the subject is advised to improve lifestyle, reduce direct sunlight, and increase indoor activity time to reduce further damage to the skin from ultraviolet rays.

[0035] Example 5: 25-year-old female subject Image data acquisition: Multi-spectral imaging is performed on a 25-year-old female subject to obtain visible light images, ultraviolet reflection images, and near-infrared spectral images. The visible light images show that the subject's facial skin is smooth with only a small number of fine lines and slight speckles; the ultraviolet reflection images reveal the fine structure of the speckle area with clear speckle edges and a small number of speckles; the near-infrared spectral images show that the skin has a high water content and sufficient collagen composition.

[0036] Image data preprocessing and feature parameter extraction: The image data is standardized, registered, and noise-removed, and the speckle area is segmented using the watershed algorithm. The speckle area ratio is calculated to be 1.0%, and the speckle density is 2.0 per square centimeter. The wrinkle distribution features are extracted using the Histogram of Oriented Gradients (HOG) method, resulting in a wrinkle distribution intensity of 0.30 (normalized value), with wrinkles mainly concentrated in the eye area. Capillary network parameters are also extracted, revealing a capillary density of 60 per square millimeter, with uniform distribution and no obvious dilation.

[0037] Three-dimensional facial skin model construction and elastic parameter acquisition: A three-dimensional scanner is used to construct a three-dimensional facial skin model of the subject. A vertical pressure of 0.5N is applied to multiple pre-selected test points on the face, and the skin deformation process is recorded by a high-frame-rate camera. Based on the viscoelastic model, the stress relaxation time constant τ of the subject's skin is calculated to be 0.8 seconds, and the elastic recovery rate η is 85%, indicating that the skin elasticity is good and the recovery ability is strong.

[0038] Deep learning aging assessment model application: The extracted feature parameters are input into the deep learning aging assessment model, and the model analyzes the parameters comprehensively through a multi-task learning architecture. The first branch network processes the texture features and color spot parameters, and obtains a texture aging index associated with age of 0.4; the second branch network processes the capillary network topology features, and obtains a vascular aging index of 0.3; the third branch network processes the elastic mechanics parameters, and obtains an elastic aging index of 0.3. Through the attention mechanism, the output features of each branch are fused, and finally the skin physiological age of the subject is obtained as 23 years old, the elasticity level is good, and the regional aging difference index is 5%.

[0039] Comprehensive evaluation report output and personalized recommendation generation: According to the evaluation results, a comprehensive evaluation report is output, which shows that the skin physiological age of the subject is lower than the actual age, the elasticity is at a good level, the facial aging degree is lighter, and mainly concentrates in the periorbital area. At the same time, individual parameters input by the subject are received, including age 25, female, living in the city, good daily skin care habits and paying attention to sun protection, etc. These individual parameters and detection parameters are input into the evaluation model together to generate personalized aging prevention suggestions, suggesting the subject to continue to maintain good skin care habits, regularly use moisturizing and antioxidant skin care products, and appropriately perform periorbital care to maintain the skin in good condition and prevent early aging.

[0040] The method and system for evaluating facial skin aging proposed in the present application realize multi-dimensional quantitative evaluation of skin aging through advanced technical means such as multi-spectral imaging, mechanical sensing and deep learning, and generate a comprehensive evaluation report containing skin physiological age, elasticity level and regional aging difference index, while providing personalized aging prevention suggestions, providing a powerful tool for skin health management. The method has the characteristics of high precision, objectivity and personalization, and can provide strong support for skin health management, and has broad application prospects. In future research, the architecture and training method of the deep learning model can be further optimized, the size and diversity of the training data set can be expanded, and the evaluation accuracy and generalization ability of the system can be improved; at the same time, more biomarkers and detection indicators related to skin aging can be explored, and the evaluation system can be continuously enriched and improved to provide more comprehensive and in-depth technical support for the research and prevention of skin aging.

Claims

1. A method of assessing facial skin aging, characterized in that Comprising the following steps: S1, acquiring multi-dimensional image data of facial skin by a multi-spectral imaging device, including visible light image, ultraviolet reflection image and near-infrared spectrum image; S2, preprocessing the multi-dimensional image data to extract skin texture feature parameters, color spot distribution parameters and capillary network parameters; S3, constructing a three-dimensional facial skin model and acquiring skin elasticity parameters through a mechanical sensor, the elasticity parameters including stress relaxation time and elastic recovery rate; S4, inputting the parameters obtained in steps S2 and S3 into an aging assessment model trained by deep learning, the model extracting multi-scale features through a convolutional neural network and comparing with an age-related database; S5, outputting a comprehensive assessment report containing skin physiological age, elasticity level and regional aging difference index.

2. The method of claim 1, wherein: The preprocessing in step S2 comprises: standardizing registration and noise elimination processing of multi-spectral images; segmenting skin color spot regions by a watershed algorithm; extracting wrinkle distribution features by a histogram of oriented gradients (HOG).

3. The method of claim 1, wherein: The acquisition of skin elasticity parameters comprises: applying a vertical pressure of 0.5-1.2 N at preset test points on the face; recording the skin deformation process through a high-frame-rate camera (≥1000 fps); calculating stress relaxation time constant τ and elastic recovery rate η based on a viscoelastic model.

4. The method of claim 1, wherein: The deep learning model adopts a multi-task learning architecture, comprising: a first branch network processing texture features and color spot parameters; a second branch network processing capillary network topology features; a third branch network processing elastic mechanics parameters; fusing output features of each branch through an attention mechanism.

5. The method of claim 4, wherein Further comprising: in the model training stage, using a transfer learning method to pre-train based on a skin multi-modal database containing 10,000 cases of different races, genders and ages; using a generative adversarial network (GAN) to enhance the training data and generate synthetic data under different lighting conditions and skin states.

6. The method of claim 1, wherein Further comprising: receiving user input individual parameters including age, gender, geographical residence information and skincare habits; inputting the individual parameters and detection parameters into the assessment model together to generate personalized aging prevention suggestions.

7. The method of claim 1, wherein: The regional aging difference index is calculated by: DI = (E_max - E_min) / E_avg × 100% where E_max is the maximum aging value in each subzone of the face, E_min is the minimum aging value, and E_avg is the average aging value.

8. An evaluation system for carrying out the method according to any one of claims 1 to 7, characterized in that Comprising: an image acquisition module containing a multi-spectral imaging device and a three-dimensional scanner; a data processing module configured with a GPU accelerated computing unit; a parameter database storing skin feature reference values based on age grouping; a result output module supporting visual charts and augmented reality (AR) simulation display.

9. The method of claim 1, wherein Further comprising: measuring skin impedance parameters through a micro-current sensor; weighting and fusing the impedance parameters with the elasticity parameters to calculate a skin metabolic activity index (MAI): MAI = α·η + β·(1 / Z_avg) where α and β are weight coefficients, and Z_avg is the average impedance value.

10. The method of claim 1, wherein: In step S5, a dynamic aging prediction curve is generated, including: Based on the current parameters, the skin state differential equation is established: dA / dt = k·(A_curr - A_base) + ε·E_loss Where A_curr is the current physiological age, A_base is the base age, E_loss is the elastic loss amount, and k, ε are adjustment coefficients.