A method and system for simulating the effects of facial minimally invasive cosmetic surgery based on generative AI.

Abstract

The application discloses a face microplasty surgery effect simulation method and system based on generative AI, relates to the technical field of face image processing, and improves the simulation precision of the face microplasty surgery effect by performing face diagnosis and processing analysis on patient face image data in combination with generative AI technology, and further generates a personalized and accurate face plastic surgery scheme in an interactive mode, so that the predictability and safety of the microplasty surgery can be significantly improved, the medical experience of the patient is optimized, the risk of the microplasty surgery is reduced, communication with the patient is facilitated, and the digitalization and intelligentization development of the microplasty industry are promoted.

Description

Technical Field

[0001] This invention relates to the field of facial image processing technology, and in particular to a method and system for simulating the effects of facial micro-plastic surgery based on generative AI. Background Technology

[0002] In the medical setting, communication between doctors and patients is crucial. As the demand for facial minimally invasive cosmetic procedures grows, consumers want to see the results in advance to make more informed decisions. Generative AI simulations can help doctors better explain facial cosmetic procedures and expected outcomes to patients, reducing misunderstandings and anxieties.

[0003] With the continuous development of modern cosmetic surgery techniques, current methods for predicting the effects of facial minimally invasive procedures mainly rely on the experience of professional doctors or some simple visual simulation tools. These traditional methods often suffer from low accuracy, difficulty in reflecting individual needs, and complex procedures, making it difficult to provide patients with precise pre-operative result predictions and personalized adjustment suggestions. Furthermore, existing techniques often fail to accurately consider the characteristics of different filler materials, the physical properties of facial skin, and the effects of light, leading to a significant discrepancy between the final post-operative results and expectations. Because different filler materials and their varying states within the patient's body make it difficult for patients to determine the degree to which the chosen material matches their individual needs.

[0004] In summary, how to combine generative AI to analyze different filler materials and their state in the patient's body to simulate and visualize the effects of minimally invasive cosmetic surgery, thereby facilitating communication with patients about the surgical plan and reducing surgical risks, is a problem that facial minimally invasive cosmetic surgery effect simulation system urgently needs to solve and optimize. Summary of the Invention

[0005] This invention provides a method and system for simulating the effects of facial minimally invasive cosmetic surgery based on generative AI. It addresses the technical problem of how to combine generative AI to analyze different filler materials and their state in the patient's body to simulate and visualize the effects of minimally invasive cosmetic surgery.

[0006] To address the aforementioned technical problems, this invention provides a method and system for simulating the effects of facial minimally invasive cosmetic surgery based on generative AI. The specific technical solution is as follows:

[0007] Firstly, a method for simulating the effects of facial minimally invasive cosmetic surgery based on generative AI includes the following steps:

[0008] Multiple types of three-dimensional facial image data of patients are collected to obtain a three-dimensional facial image dataset; facial defect features of each data item in the three-dimensional facial image dataset are extracted and labeled to obtain a defect labeling dataset, wherein the defect labeling data represents the marking of defect feature data existing on the patient's face;

[0009] Based on the historically collected dataset of three-dimensional facial images of patients and the defect calibration dataset, a generative AI model is used to identify the defect features in the three-dimensional facial images of patients and simulate an initial micro-plastic surgery plan based on the defect features of the patients' faces.

[0010] Based on the initial minimally invasive cosmetic procedure, patients can intuitively experience the comparison of facial improvement before and after the surgery. Based on the comparison results and real-time interactive feedback with patients, the surgical details are adjusted multiple times according to the patients' expected results to set real-time adjustment strategies and generate the final facial minimally invasive cosmetic procedure.

[0011] Based on the final facial micro-plastic surgery plan, the interactive effect model of facial micro-plastic surgery is constructed to obtain the interactive effect of facial micro-plastic surgery filling or dissolving materials with the human face, and the generative AI model is updated and fed back to obtain the final visual display image of the face after surgery.

[0012] As a further optimization of the present invention, facial image data of the patient is scanned and collected, and three-dimensional modeling of the patient's facial image is performed to generate three-dimensional facial image data; multiple facial three-dimensional image data are combined into a facial three-dimensional image dataset; and two-dimensional image data in the four directions of top, bottom, left and right of each facial three-dimensional image are retained to obtain a micro-plastic surgery reference image dataset.

[0013] As a further optimization of the present invention, the real-time adjustment strategy includes:

[0014] The initial minimally invasive surgical plan is based on the patient's facial defect feature data to obtain the degree of facial defect; a preset facial smoothness threshold range is set, and the degree of facial defect is compared with the preset facial smoothness threshold range. Based on the comparison result, the degree of facial defect is determined to be either concave or convex.

[0015] pass To obtain the degree of protrusion or concavity of facial defects, where R1 and R2 represent the principal curvature radii of the protruding or concave regions, respectively; when facial depression features appear, the facial depression area is filled with filler material to make the depression area reach the preset facial smoothness threshold range; when facial protrusion features appear, the facial protrusion area is dissolved with dissolving material to make the protrusion area reach the preset facial smoothness threshold range, so as to obtain a secondary micro-plastic surgery solution.

[0016] As a further optimization of the present invention, based on the aforementioned minimally invasive secondary scheme, and according to the patient's feedback on satisfaction with the surgical results, description of facial changes, and personal preferences, the updated generative AI model is used to output the adjustment simulation results.

[0017] The adjustment simulation results include adjustments to facial contours, skin firmness, fat distribution, and filling depth to dynamically optimize the patient's facial harmony in multiple dimensions. Based on the adjustment simulation results, multiple adjustment schemes are recommended through self-recommendation to achieve a final personalized facial micro-plastic surgery plan, and iterative optimization is carried out in combination with individual patient differences to obtain a three-level micro-plastic surgery plan.

[0018] As a further optimization of the present invention, facial defect features of each data item in the facial three-dimensional image dataset are extracted and labeled to obtain a defect labeling dataset, including:

[0019] The coordinates of each data item in the facial 3D image dataset are standardized to a uniform range through centering and scaling to obtain a normalized 3D image dataset. The 3D image model of the normalized 3D image dataset is aligned by the iterative nearest point algorithm so that different facial images are in the same reference coordinate system.

[0020] Defect regions are identified by statistical analysis of the three-dimensional shape of the face. Key points are marked and positioned in the three-dimensional image data items in the normalized three-dimensional image dataset. Multiple key points are arranged around the facial defect region. Facial defect region features are extracted by three-dimensional point cloud technology to obtain facial defect region features.

[0021] Based on facial defect region features extracted from 3D images, a sliding window algorithm is used to find defect regions and perform calibration processing to obtain defect calibration data; multiple defect calibration data are obtained from various different 3D facial images to form a defect calibration dataset.

[0022] As a further optimization of the present invention, the characteristic data of micro-plastic surgery filling or dissolving materials that conform to human tissue are obtained to obtain a material characteristic dataset; the historically collected facial three-dimensional image dataset and the material characteristic dataset are used to construct a training set to generate a facial micro-plastic surgery interaction effect model; the facial micro-plastic surgery interaction effect model outputs the interaction effect recognition result of the micro-plastic surgery filling or dissolving materials required for the facial image.

[0023] The newly acquired 3D image of the patient's face is input into the facial micro-plastic surgery interaction effect model to output the facial defect features of the newly acquired 3D image of the patient's face and the prediction results of the interaction effect of the required filling or dissolving materials; based on the interaction effect prediction results, the dynamic change characteristics of the filling or dissolving materials in the patient's body are used to obtain the interaction data between the simulated materials and facial tissues.

[0024] As a further optimization of the present invention, based on the recognition results of the facial micro-plastic surgery interaction effect model, and combined with the characteristic data of micro-plastic surgery filling or dissolving materials, a simulation is performed; according to the simulated effect of the filling or dissolving material interaction effect on the patient's face, a postoperative effect visualization image is generated to show the improvement effect of the filling or dissolving materials on facial defect features; based on the displayed effect, feedback is provided in real time and the plan is adjusted in a timely manner to generate a personalized facial micro-plastic surgery final plan.

[0025] As a further optimization of the present invention, based on the characteristic data of different filling or dissolving materials, the deformation effect of the filling or dissolving materials is obtained by using σ=E·ε; where σ represents stress, E represents the elastic modulus of the material characteristic data, and ε represents strain;

[0026] Based on the deformation effect of the filling or dissolving material, through The diffusion degree of the filling or dissolving material in the subcutaneous tissue gap is obtained to describe the adaptation and distribution of the material on the face; and combined with the facial micro-plastic surgery interactive effect model, the effect of the material filling or dissolving on the facial surface morphology is obtained.

[0027] In the formula, δ represents the diffusion degree of the filling or dissolving material in the subcutaneous tissue interstitial space; F represents the deformation force generated by the filling or dissolving material; k skin This represents the skin's elasticity coefficient.

[0028] As a further optimization of the present invention, based on the influence of the facial surface morphology after filling or dissolving, B-spline deformation is used to smooth the facial lines of the facial micro-plastic 3D image generated by the generative AI model.

[0029] by I=I t +I d +I s To obtain a realistic representation of the post-operative lighting effect on the face and to map facial textures to improve the realism of the visualization; the facial textures include skin tone uniformity, texture coherence, and smoothness; where, I t I d and I s These represent the brightness of ambient light, diffuse light, and specular light, respectively, while I represents the postoperative illumination of the face as actually displayed.

[0030] The visualization image is used to view the effect of the micro-plastic surgery, and to adjust various parameters such as the amount of filling material, its position, and its dissolution effect; and to update the generative AI model to generate the final visualized image of the postoperative effect of the facial micro-plastic surgery.

[0031] Secondly, the system includes an electronic device comprising a memory, a processor, and a generative AI-based facial minimally invasive cosmetic surgery effect simulation method program stored in the memory and executable on the processor. When executed by the processor, the generative AI-based facial minimally invasive cosmetic surgery effect simulation method program implements the steps of a generative AI-based facial minimally invasive cosmetic surgery effect simulation method. The system includes:

[0032] Data acquisition module: It is used to acquire various three-dimensional facial image data of patients to obtain a three-dimensional facial image dataset; extract and label the facial defect features of each data item in the three-dimensional facial image dataset to obtain a defect labeling dataset, wherein the defect labeling data represents the marking of defect feature data existing on the patient's face;

[0033] Data recognition module: It is used to identify the defect feature data of the patient's facial three-dimensional image dataset and the defect labeling dataset collected in history, and to simulate the initial micro-plastic surgery plan based on the patient's facial defect feature data through a generative AI model;

[0034] Real-time adjustment module: It is used to allow patients to intuitively feel the comparison effect of facial improvement before and after surgery based on the initial micro-plastic surgery plan; based on the comparison results and real-time interactive feedback with the patient, it adjusts the surgical details in multiple rounds according to the patient's expected effect, so as to set the real-time adjustment strategy and generate the final facial micro-plastic surgery plan;

[0035] Effect Display Module: Based on the final facial micro-plastic surgery plan, it constructs the facial micro-plastic surgery interactive effect model to obtain the interaction effect between the facial micro-plastic surgery filling or dissolving materials and the human face, and updates the generative AI model to obtain the final visual display image of the face after surgery.

[0036] The present invention offers at least the following beneficial effects: By acquiring three-dimensional facial image data from patients, the present invention can obtain detailed information such as facial structure, skin texture, and skeletal features. This data foundation provides accurate and detailed information for subsequent analysis and processing; each patient has a different facial shape, and three-dimensional data acquisition can help generate personalized facial models for each patient, laying a solid foundation for precise planning of minimally invasive cosmetic surgery; acquiring three-dimensional data from different patients and establishing diverse datasets can improve the generalization ability of the model, enabling the AI ​​to adapt to the facial features of different populations and avoiding the impact of data bias from individual patients.

[0037] Through feature extraction and labeling, specific facial defects (such as wrinkles, facial asymmetry, sagging, and fat accumulation) can be accurately identified and located. Labeled datasets provide a clear basis for subsequent defect correction and minimally invasive cosmetic procedures. These datasets offer consistent standards for different doctors and AI models, contributing to a unified diagnostic and treatment reference. They can also serve as "label data" for training generative AI models, improving model accuracy. Quantitative labeling of defects allows doctors to better understand the problems on a patient's face, enabling them to develop minimally invasive cosmetic procedures that better meet individual needs.

[0038] Generative AI models can automatically identify facial defects in patients based on historical datasets and simulate possible corrective measures, greatly improving work efficiency and accuracy. The preliminary plans generated by AI simulations can take into account each patient's unique facial features, such as facial bone structure, muscle movement, and skin tension, making the initial plans more in line with the patient's individual needs. The application of AI models helps reduce subjective errors in the surgical planning process, providing objective plans based on a large amount of data and model reasoning. The generated initial plans can be visualized to show the postoperative effects to the patient, allowing the patient to intuitively feel the possible changes after surgery and increasing their confidence in the surgical plan.

[0039] Through real-time interaction between patients and the AI ​​system, the AI ​​can adjust surgical plans based on patient feedback to ensure postoperative results meet patient expectations. For example, a patient might desire a more natural or three-dimensional appearance in a particular area, which the AI ​​can dynamically adjust. This interaction increases patient participation in the decision-making process, enabling customized adjustments to the surgical plan and increasing patient satisfaction. Patient feedback can be used for multiple rounds of adjustments, allowing the generative AI model to accumulate experience in each round and gradually refine the surgical plan through algorithm optimization, making it more tailored to patient needs. By continuously comparing patient expectations with model output, real-time before-and-after facial comparisons can be displayed, helping patients make informed decisions and enhancing their confidence in the surgery.

[0040] The interactive effect model accurately simulates the behavior, interactions, and impact on facial contours of filler or dissolving materials on the skin. By simulating the effects of filling, dissolving, or other minimally invasive cosmetic procedures, it can better predict post-operative results. The model considers the interaction between the material and the skin, such as how the filler material forms support under the skin and how muscles affect the material's flow, thus more realistically reflecting the post-operative effect. Through real-time feedback and adjustments, the AI ​​model continuously optimizes its predictions and effect simulations. This closed-loop mechanism ensures that the generated final plan highly matches the patient's expected results.

[0041] Visualizing postoperative results helps patients clearly understand the possible facial changes after surgery, enabling them to better comprehend the potential benefits of the procedure. By comparing their actual results with their expected results using these visual images, patients can make more informed decisions. Precise preoperative simulations and visualizations can significantly reduce the uncertainty of postoperative outcomes, helping patients set better expectations and avoid unnecessary psychological stress after surgery. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of a method for simulating the effect of facial micro-plastic surgery based on generative AI, provided by an embodiment of the present invention;

[0043] Figure 2 This is a schematic diagram of a facial micro-plastic surgery effect simulation system based on generative AI, provided by an embodiment of the present invention. Detailed Implementation

[0044] The present application will now be described in further detail with reference to the accompanying drawings. It should be noted that the following specific embodiments are only used to further illustrate the present application and should not be construed as limiting the scope of protection of the present application. Those skilled in the art can make some non-essential improvements and adjustments to the present application based on the above application content.

[0045] This embodiment provides a method and system for simulating the effects of facial minimally invasive cosmetic surgery based on generative AI. The specific implementation method is as follows:

[0046] like Figure 1 As shown, a method for simulating the effects of facial minimally invasive cosmetic surgery based on generative AI includes the following steps:

[0047] Step 11: Collect various three-dimensional facial image data of patients to obtain a three-dimensional facial image dataset; extract and label the facial defect features of each data item in the three-dimensional facial image dataset to obtain a defect labeling dataset, wherein the defect labeling data represents the labeling of defect feature data existing on the patient's face;

[0048] Step 12: Based on the historically collected patient facial 3D image dataset and the defect calibration dataset, a generative AI model is used to identify the defect feature data of the patient's facial 3D images and simulate an initial micro-plastic surgery plan based on the patient's facial defect feature data.

[0049] Step 13: Based on the initial minimally invasive cosmetic surgery plan, allow the patient to intuitively experience the comparison effect of facial improvement before and after the surgery; based on the comparison results and real-time interactive feedback with the patient, adjust the surgical details in multiple rounds according to the patient's expected effect, so as to set a real-time adjustment strategy and generate the final facial minimally invasive cosmetic surgery plan.

[0050] Step 14: Based on the final facial micro-plastic surgery plan, construct the facial micro-plastic surgery interactive effect model to obtain the interaction effect between the facial micro-plastic surgery filling or dissolving materials and the human face, and update the generative AI model to obtain the final visual display image of the face after surgery.

[0051] In this invention, step 11 involves collecting facial 3D image data from various patients to establish a comprehensive facial 3D dataset. This step not only collects data but also includes acquiring facial images from different angles and with different expressions. This aims to ensure the diversity and representativeness of the dataset, better covering various possible facial features and defect types. By extracting and labeling the defect features of each patient's face, a defect labeling dataset is formed. This labeling process refers to clearly marking and classifying each facial defect, including but not limited to facial depressions, protrusions, wrinkles, sagging, and looseness, thereby providing accurate basic data for subsequent analysis; thus establishing a large and diverse database containing data on different types of facial defects. The diversity of the data provides a more comprehensive foundation for subsequent AI model training, enhancing its generalization ability and accuracy.

[0052] Step 12, based on the facial 3D image dataset and defect labeling dataset collected in Step 11, employs a generative AI model for training and analysis. The core objective of this step is to accurately identify the facial defect features of each patient using historical data and defect labeling information, and to generate an initial minimally invasive cosmetic surgery plan. By analyzing and learning from this data, the generative AI model can predict and simulate a suitable minimally invasive cosmetic surgery plan for the patient. The AI ​​model not only simply identifies defects but also considers factors such as the patient's overall facial structure and proportions to propose a preliminary surgical plan; this significantly improves the efficiency and accuracy of plan development. The AI's analytical capabilities can identify details that are difficult to detect with the naked eye, ensuring the rationality and personalization of the initial plan. Furthermore, the generated plan is more scientifically sound and data-driven, providing a strong basis for subsequent adjustments.

[0053] Step 13, based on the initial minimally invasive cosmetic surgery plan generated in Step 12, uses efficient visualization tools (such as 3D model display, virtual reality, etc.) to simulate the facial improvement before and after surgery, allowing the patient to more intuitively experience the results. During this process, the patient can interact with the doctor in real time, providing feedback on their expectations and requirements for the surgical outcome. Based on the patient's feedback, the doctor can make multiple adjustments, including modifying facial proportions and adjusting filler areas, ultimately arriving at a final minimally invasive facial cosmetic surgery plan that satisfies the patient. This process emphasizes interaction with the patient, ensuring the final plan best meets their needs. Through real-time feedback and multiple adjustments, this stage allows the patient to better understand the post-operative results and enhances their trust and satisfaction with the surgical plan. Through visualization simulation, patients can intuitively see their potential changes before surgery, avoiding the anxiety and misunderstandings caused by the inability to accurately predict results in traditional methods.

[0054] Step 14: After the final minimally invasive cosmetic procedure plan is determined in Step 13, an interactive effect model for facial minimally invasive cosmetic procedures is constructed. This model not only considers the interaction between facial filler or dissolving materials and the human face, but also incorporates changes during the postoperative recovery period. Through feedback and updating of the generative AI model, a final visual image of the face after surgery can be obtained. Using this model, patients can anticipate the postoperative results, including not only morphological changes but also the long-term impact of factors such as the biocompatibility and durability of the surgical materials on the face; thus ensuring the long-term predictability and accuracy of the surgical outcome. By constructing the interactive effect model, some changes and potential problems during the postoperative recovery process can be predicted in advance, helping patients better understand the long-term postoperative results. Simultaneously, updates to the AI ​​model can continuously optimize and improve the facial minimally invasive cosmetic procedure plan, making it more suitable for individual differences and biological characteristics.

[0055] The four steps outlined above form a precise and systematic facial minimally invasive cosmetic surgery design process. From data collection and AI model generation to interactive adjustments with the patient, and finally to optimization of the interactive effect model, each step is interconnected and mutually supportive. This not only significantly improves the precision and personalization of minimally invasive cosmetic surgery but also enhances the patient's experience and sense of trust through visualization and interaction. The beneficial effects of each step complement each other, improving the overall quality of medical services and patient satisfaction, while also making the minimally invasive cosmetic surgery process more scientific, transparent, and controllable.

[0056] In a preferred embodiment of the present invention, step 11 further includes:

[0057] Step 111: Scan and collect facial image data of the patient, perform three-dimensional modeling on the patient's facial image to generate facial three-dimensional image data; combine the various facial three-dimensional image data into the facial three-dimensional image dataset; and retain the two-dimensional image data of each facial three-dimensional image in the four directions of top, bottom, left and right to obtain the micro-plastic surgery reference image dataset.

[0058] In this invention, step 111 first involves scanning and acquiring facial image data of the patient. Facial image data is acquired using professional 3D scanning equipment (such as an optical scanner or structured light camera). This process captures detailed facial information from multiple angles and directions, ensuring the scanned data is sufficiently accurate and complete. The acquired data typically includes facial geometry, surface details, and texture features. Next, advanced image processing techniques are used to create a 3D model of the acquired facial image data. This modeling process utilizes computer vision and 3D reconstruction algorithms to integrate image data from different angles, generating a comprehensive and accurate 3D facial model. This 3D model not only reflects the geometric shape of the face but also contains detailed surface texture information, such as skin texture and minute details.

[0059] After completing the 3D modeling, the collected 3D image data are aggregated to form a facial 3D image dataset. This dataset contains facial 3D data from different patients, covering a variety of facial features and structural variations. To further improve the diversity and representativeness of the data, the procedure also requires retaining the 2D image data in the top, bottom, left, and right directions of each facial 3D image. These 2D images will help to more accurately capture various angles and local features of the face, thus supplementing the detailed information in the 3D modeling. All of these data are aggregated to form a minimally invasive cosmetic surgery reference image dataset.

[0060] Multiple facial images from various angles are acquired using 3D scanning equipment to ensure sufficient detail is captured. This image data includes facial features observed from different perspectives (such as front, side, top, and bottom) to comprehensively reconstruct the 3D structure of the face. These 2D image data are then processed using 3D reconstruction algorithms to generate high-precision 3D facial models, preserving detailed geometry, curves, and textures. This process relies on computer vision techniques and image fusion algorithms to ensure high fidelity of the facial models. Facial 3D image data from various patients are aggregated into a single dataset, and 2D image data from different angles for each 3D model is also saved, further increasing the diversity and richness of the dataset. By maintaining 2D images from four directions, the dataset can provide a more comprehensive reference for subsequent processing.

[0061] By employing 3D scanning technology and multi-angle image acquisition, facial details can be precisely captured, generating highly realistic 3D facial models. This significantly improves the accuracy of subsequent facial minimally invasive cosmetic surgery design, avoiding errors caused by traditional 2D images or single-view data. The retained 2D image data from four directions provides a more comprehensive reference for subsequent facial defect analysis and minimally invasive surgery design. These 2D images complement the details of the 3D model, enhancing the multidimensionality and accuracy of the data.

[0062] A rich facial dataset was constructed by aggregating 3D facial image data from various patients. This dataset not only covers a wide range of facial feature variations but also provides ample sample data for training AI models, facilitating the generation of personalized and precise minimally invasive cosmetic surgery plans. The establishment of the 3D facial image dataset and its 2D reference images lays the foundation for subsequent work such as facial defect feature extraction, AI model training, and minimally invasive cosmetic surgery plan design. By using this dataset, doctors and AI systems can more accurately identify the individualized features and defects of patients' faces, and formulate personalized and expected minimally invasive cosmetic surgery plans.

[0063] By scanning and acquiring patients' facial images and creating 3D models, combined with 2D image data from four directions, a detailed and accurate 3D facial dataset was established. This dataset not only provides high-quality foundational data for designing facial minimally invasive cosmetic procedures but also ensures data comprehensiveness and diversity, contributing to improved accuracy and personalization of the procedure. The beneficial effect of this process lies in its strong support for subsequent defect identification, AI model generation, and postoperative outcome prediction.

[0064] In a preferred embodiment of the present invention, the real-time adjustment strategy in step 13 includes:

[0065] Step 131: Based on the patient's facial defect feature data, a micro-plastic surgery initial plan is developed to obtain the patient's facial defect degree; a preset facial smoothness threshold range is established, and the patient's facial defect degree is compared with the preset facial smoothness threshold range. Based on the comparison result, the patient's facial defect degree is determined to be either concave or convex.

[0066] Step 132, through To obtain the degree of protrusion or concavity of facial defects, where R1 and R2 represent the principal curvature radii of the protruding or concave regions, respectively; when facial depression features appear, the facial depression area is filled with filler material to make the depression area reach the preset facial smoothness threshold range; when facial protrusion features appear, the facial protrusion area is dissolved with dissolving material to make the protrusion area reach the preset facial smoothness threshold range, so as to obtain a secondary micro-plastic surgery solution.

[0067] In this invention, step 131 involves developing an initial minimally invasive cosmetic procedure based on the patient's facial defect features (e.g., facial depressions, protrusions, wrinkles, etc.). By analyzing facial image data, defective areas are identified, and a preliminary adjustment plan is developed based on this data. This defect feature data can be obtained through 3D modeling and algorithmic analysis to ensure the accuracy of the defect information. Next, a preset facial smoothness threshold range is used for evaluation. The facial smoothness threshold is a standard range used to measure whether different areas of the face have achieved the expected smoothness. This threshold is typically set based on different facial region characteristics, facial aesthetic standards, and the patient's individual needs.

[0068] By comparing the patient's facial defects to a preset facial smoothness threshold range, the system can determine whether the defective areas are concave or convex. This comparison process helps to further analyze the overall condition of the patient's face, providing a basis for subsequent treatment plans. Through 3D scanning and image processing technology, the system acquires facial defect data, which typically includes features such as concave areas, convex areas, and uneven skin. The facial smoothness threshold serves as a standard to determine whether facial areas meet aesthetic or clinical requirements. By comparing the patient's facial defects with this threshold, it is possible to clearly identify which areas require further minimally invasive cosmetic treatments (such as filling or dissolving). Based on the comparison with the smoothness threshold, facial defects are classified as either concave or convex. This allows for personalized treatment planning for different types of facial defects.

[0069] Skin smoothness, including the overall smoothness of facial skin, wrinkles, and depressions;

[0070] The smoothness of facial contours, such as the degree of unevenness in areas like the chin, cheekbones, and forehead.

[0071] Skin smoothness is usually expressed as the roughness of the subcutaneous tissue gaps. Common measurement methods include the acquisition and analysis of skin texture using subcutaneous tissue gap roughness analyzers (such as dermoscopy or 3D scanners).

[0072] Good smoothness, with skin roughness in the range of 0.5μm-1.0μm;

[0073] Medium smoothness, with skin roughness in the range of 1.0μm-1.5μm;

[0074] Poor smoothness, with skin roughness exceeding 1.5μm.

[0075] Facial contour smoothness can be analyzed using facial scanning and 3D reconstruction techniques, and is usually quantified by changes in facial curvature.

[0076] Good flatness, with facial curvature variation ≤1.0mm.

[0077] Medium flatness, with facial curvature variation between 1.0mm and 2.0mm.

[0078] Poor flatness, facial curvature variation **>2.0mm**.

[0079] The degree of facial depressions / protrusions, especially deficiencies or prominences in areas such as the bridge of the nose, cheekbones, and forehead.

[0080] Signs of skin aging, such as wrinkles and sagging;

[0081] Pigmentation or acne scars, skin pigmentation or scars.

[0082] The height difference of different areas of the face is measured using a 3D facial scanning device.

[0083] Low defect rate, with the height difference in the facial area not exceeding 1.0 mm.

[0084] Medium defect level, with a height difference of 1.0mm-2.0mm in the facial area.

[0085] High defect rate, with a height difference in the facial area exceeding 2.0 mm.

[0086] Skin aging can be assessed by assessing skin elasticity, fine lines, and sagging. Commonly used parameters include the Elasticity Index and the number of fine lines.

[0087] Low defect rate: Skin elasticity index greater than 70, no obvious fine lines;

[0088] Moderate defect level: Skin elasticity index 50-70, a few fine lines;

[0089] High defect rate: Skin elasticity index less than 50, with obvious fine lines and sagging.

[0090] The need to adjust a minimally invasive cosmetic procedure can be determined by comparing the values ​​of smoothness and defect severity. For example, a patient's facial skin roughness is 1.2μm (moderate smoothness), facial curvature variation is 1.8mm (moderate smoothness), and the main facial defects are under-eye bags (2.5mm depth) and fine lines on the forehead (skin elasticity index 55). Since both skin and contour smoothness are within the moderate range, improvement in skin smoothness and contour symmetry is needed. The under-eye bags and forehead fine lines indicate moderate facial defects, requiring localized filling and tightening treatments within the minimally invasive procedure plan.

[0091] In another example, another patient had a skin roughness of 0.8 μm (good smoothness) and a facial curvature variation of 0.5 mm (good smoothness), but had significant cheekbone prominence (height difference of 2.2 mm) and neck laxity (skin elasticity index of 45). The smoothness of the skin and contour was good, but local areas (cheekbones and neck) required special attention. The height difference of the cheekbones was large, and the neck laxity was more obvious, requiring adjustment through filling and tightening measures.

[0092] Step 132, based on the analysis results from Step 131, determines the specific type of facial defect (depression or protrusion). Next, using mathematical formulas or algorithmic models, the degree of depression or protrusion in the defective area is further calculated. In this process, the principal radii of curvature (R1 and R2) are used to describe the curvature of the depressed or protruding facial area; R1 and R2 represent the principal radii of curvature of the depressed or protruding area, respectively, used to quantify the curvature of the defective area. By calculating these principal curvature values, the severity of the depressed or protruding facial area is accurately assessed. The purpose of this step is to provide a more precise minimally invasive cosmetic procedure to ensure optimal treatment results.

[0093] Based on this analysis, if facial depressions are present, they are typically filled with fillers to bring the depressed areas to a predetermined facial smoothness threshold. Fillers such as hyaluronic acid and collagen effectively restore the smoothness of depressed areas and have good biocompatibility. If facial protrusions are present, dissolving materials are used. Dissolving materials (such as dissolving fillers) help dissolve or reduce the volume of protruding areas, smoothing them and restoring a suitable facial contour. The ultimate goal is to provide targeted repair for the patient's facial depressions or protrusions based on the facial smoothness threshold, developing a minimally invasive secondary cosmetic procedure that conforms to facial aesthetics.

[0094] By obtaining the principal radii of curvature (R1 and R2) of the defective area, the curvature changes of the facial region are quantified, thereby accurately assessing the severity of the defect. This calculation helps determine which parts need filling or dissolving; for sunken areas, appropriate filler materials, such as hyaluronic acid or collagen, are selected to fill these areas to restore smoothness; for protruding areas, dissolving materials are used to dissolve them, thereby restoring the facial contour to an ideal state. Dissolving materials help reduce excessive protrusion and restore natural facial lines; by calculating the degree of protrusion and concavity of facial defects and selecting appropriate materials for filling or dissolving, the facial contour can be precisely adjusted to achieve the desired aesthetic effect; through specific analysis of facial depressions and protrusions, the treatment process becomes more targeted and controllable, reducing uncertainties during treatment; personalized minimally invasive cosmetic plans are customized according to the specific defect characteristics of the patient's face, ensuring that each patient achieves the best treatment results. By precisely calculating the degree of unevenness, the treatment plan is more accurate, avoiding overtreatment or undertreatment; by accurately applying filling and dissolving materials, it is possible to restore facial smoothness while maintaining the natural beauty of the face, avoiding unnatural or artificial effects, and improving the naturalness and durability of the postoperative results.

[0095] In a preferred embodiment of the present invention, step 13 further includes:

[0096] Step 133: Based on the aforementioned minimally invasive secondary plan, and according to the patient's feedback on satisfaction with the surgical results, description of facial changes, and personal preferences, the updated generative AI model is used to output the adjustment simulation results.

[0097] Step 134: The adjustment simulation results include adjustments to facial contours, skin tightness, fat distribution, and filling depth to dynamically optimize the patient's facial harmony in multiple dimensions. Based on the adjustment simulation results, multiple adjustment schemes are recommended through self-recommendation to obtain a final personalized facial micro-plastic surgery plan, and iterative optimization is carried out in combination with individual patient differences to obtain a three-level micro-plastic surgery plan.

[0098] In this invention, the core of step 133 is to use patient feedback (including satisfaction with surgical results, needs for facial changes, personal preferences, etc.) as input data, combined with existing minimally invasive secondary surgical plans, and analyze and adjust it using an updated generative AI model. This generative AI model can simulate and predict the specific impact of different adjustment plans on the patient's face, and then output the adjusted simulation results. It obtains the patient's evaluation and expectations of the existing surgical results through questionnaires, dialogue boxes, or direct communication, including needs regarding facial contours, skin texture, fat distribution, etc. Based on this feedback and existing secondary plan data, the AI ​​model performs calculations and simulations to predict which adjustment measures can most effectively meet the patient's needs. Through this simulation result, it can provide doctors with optimized facial plans and assist patients in understanding the adjusted effects. By collecting and analyzing specific patient feedback, it ensures that the adjusted plan better matches the patient's individual needs and expectations. The AI ​​model can more accurately predict the impact of different adjustments on the patient's facial results, reducing surgical risks and uncertainties.

[0099] Step 134 further optimizes multiple dimensions, including facial contour, skin firmness, fat distribution, and filling depth, based on the simulation results from Step 133. The system optimizes the harmony of the patient's face according to changes in these multiple dimensions, ultimately generating multiple minimally invasive cosmetic surgery plans and recommending them to the patient for selection. Adjusting the simulation results does not only focus on a single facial feature but also considers the overall harmony of the face, dynamically optimizing multiple factors such as facial contour, skin firmness, fat distribution, and filling depth. Based on individual patient differences (such as facial bone structure, skin type, and facial muscle condition), the AI ​​system can continuously iterate and optimize the minimally invasive cosmetic surgery plan, proposing a series of different adjustment options for the patient to choose from. Through continuous optimization of the adjustment plan, it can automatically provide the patient with the final personalized minimally invasive cosmetic surgery plan that best suits their needs.

[0100] Through multi-dimensional optimization, patients achieve more natural and harmonious facial results, avoiding abrupt adjustments in certain areas; precise recommendations are made based on individual patient differences to ensure that the final minimally invasive cosmetic procedure meets the patient's unique facial structure and aesthetic needs; and continuous optimization based on feedback avoids repeated adjustments or dissatisfaction caused by inappropriate procedure selection in traditional surgery.

[0101] In a preferred embodiment of the present invention, step 11 involves extracting and labeling the facial defect features of each data item in the facial three-dimensional image dataset to obtain a defect labeling dataset, including:

[0102] Step 112: The coordinates of each data item in the three-dimensional facial image dataset are standardized to a uniform range through centering and scaling to obtain a normalized three-dimensional image dataset; the three-dimensional image model of the normalized three-dimensional image dataset is aligned by the iterative nearest point algorithm so that different facial images are in the same reference coordinate system.

[0103] Step 113: Identify defect areas through statistical analysis of the three-dimensional shape of the face; mark key points for three-dimensional positioning of the three-dimensional image data items in the normalized three-dimensional image dataset; arrange multiple key points around the facial defect area; and extract facial defect area features through three-dimensional point cloud technology to obtain facial defect area features.

[0104] Step 114: Based on the facial defect region features extracted from the 3D image, a sliding window algorithm is used to find the defect region and perform calibration processing to obtain defect calibration data; multiple defect calibration data are obtained from various different 3D facial images to form a defect calibration dataset.

[0105] In this invention, step 112 involves centering and scaling the 3D facial image dataset to standardize the data. Centering involves translating all coordinates of the 3D dataset so that the center of the dataset aligns with the origin of the coordinate system. Then, scaling adjusts all coordinates proportionally, ensuring that the coordinates of the data items fall within a uniform range of [-1, 1]. This standardization helps eliminate scale and positional differences between different facial datasets. After standardization, different facial images can be mapped to the same reference coordinate system, eliminating geometric differences caused by factors such as shooting angle and distance. Standardized facial images, operating in the same coordinate system, allow for further comparison and analysis, avoiding errors caused by data inconsistency. This provides a good foundation for the subsequent Iterative Closest Point (ICP) algorithm, enabling precise alignment of different facial images and ensuring the accuracy of subsequent analysis.

[0106] Step 113 identifies potential defect areas through statistical analysis of the three-dimensional shape of the face. First, key points are established using image marker 3D localization on the normalized 3D image data. This image marker 3D localization means locating facial defect feature markers on the three-dimensional facial images. These key points are distributed across the facial 3D dataset, typically surrounding areas where defects may exist (such as eye bags, nasolabial folds, etc.). These key points are used to mark the geometric features of different facial regions, especially in areas of potential defects. Next, 3D point cloud technology is used to further analyze the facial data around the key points to extract defect area features. Point cloud technology, through detailed analysis of facial 3D data, can capture minute geometric differences, such as skin depressions, swelling, or other defect features. By marking key points in 3D and arranging them around the defect area using image markers, the location of facial defects can be accurately pinpointed, ensuring that subsequent analysis focuses on the problem area. Point cloud technology can capture minute facial deformations, which helps to identify and quantify facial defects, such as wrinkles, depressions, or asymmetry, thus providing data support for defect repair. By extracting features from the defect area, precise input can be provided for subsequent repair algorithms, ensuring natural and accurate repair results.

[0107] Step 114 further calibrates the defect areas in the facial image using the sliding window algorithm. The sliding window algorithm is an algorithm that moves progressively across a facial image. It can automatically locate and calibrate defect areas, precisely pinpointing the specific location and shape of each facial defect through scanning multiple windows. Each window is used to calibrate the defect and generate a set of defect calibration data, i.e., the specific location, shape, and size of each defect area. By processing multiple different 3D facial images, multiple defect calibration datasets can be obtained and aggregated into a single defect calibration dataset. This dataset contains defect features from different individuals and can be used for subsequent personalized facial repair and model training. The sliding window algorithm can automatically identify and calibrate defect areas without manual intervention, improving the efficiency and accuracy of calibration. It can capture subtle differences in facial defects, generating accurate defect data to ensure the efficiency and effectiveness of subsequent repair. By processing different facial images, diverse defect calibration data is aggregated, providing data support for subsequent personalized repair schemes and enabling the training and optimization of machine learning models, thus improving the intelligence level of facial repair technology.

[0108] In a preferred embodiment of the present invention, step 14 includes:

[0109] Step 141: Obtain the characteristic data of micro-plastic surgery filling or dissolving materials that conform to human tissue to obtain a material characteristic dataset; construct a training set by combining the historically collected facial 3D image dataset and the material characteristic dataset to generate a facial micro-plastic surgery interaction effect model; the facial micro-plastic surgery interaction effect model outputs the interaction effect recognition result representing the micro-plastic surgery filling or dissolving materials required for the facial image.

[0110] Step 142: Input the newly acquired 3D image of the patient's face into the facial micro-plastic surgery interaction effect model to output the facial defect features of the newly acquired 3D image of the patient's face and the prediction results of the interaction effect of the required filling or dissolving materials; based on the interaction effect prediction results, obtain the interaction data between the simulated materials and facial tissues by using the dynamic change characteristics of the filling or dissolving materials in the patient's body.

[0111] In this invention, step 141 involves collecting data on the characteristics of micro-plastic surgery fillers or dissolving materials that are compatible with human tissue. This data includes information on the physicochemical properties of different fillers or dissolving materials, their compatibility with skin tissue, absorption rate in vivo, and elasticity changes. These datasets are built based on historically collected samples, covering the characteristics of various commonly used micro-plastic surgery materials (such as hyaluronic acid and botulinum toxin). Then, the facial 3D image dataset is combined with the material characteristic dataset to construct a training set. This training set is used to generate a facial micro-plastic surgery interactive effect model. This model can predict the effect of the required micro-plastic surgery filler or dissolving material based on the input 3D facial image, that is, predict the interactive effect of facial defect repair through the action of different materials on the face.

[0112] By combining facial 3D data and material property data, the model can more accurately predict the effects of minimally invasive cosmetic procedures, reducing side effects or repair failures caused by inappropriate material selection. The accumulation and training of historical data enable the model to have personalized customization capabilities, generating suitable minimally invasive filling plans for each patient's specific facial features and defects. The model can provide doctors or plastic surgeons with a data-based prediction tool to help them better understand the impact of different materials on facial morphology and optimize treatment plans.

[0113] Step 142 inputs a newly acquired 3D image of the patient's face. By inputting this image into the previously generated facial micro-plastic surgery interaction model, the model outputs predictions of the interaction effects of the patient's facial defects and the required filler or dissolving materials. The predictions not only identify facial defects but also calculate the appropriate type of micro-plastic surgery material and its specific application amount based on the model. Subsequently, based on these predictions, the dynamic changes of the filler or dissolving materials within the patient's body are further analyzed. This step simulates the interaction between these materials and facial tissues, taking into account factors such as material absorption, metabolism, and biocompatibility, to obtain detailed data on the changes of the materials in vivo.

[0114] By inputting a patient's 3D facial image, the model can provide a personalized minimally invasive filling or dissolving solution. It can accurately identify facial defects and tailor the most suitable treatment plan. Simulating the interaction between materials and facial tissue helps doctors predict the actual effects of different materials in the patient's body, such as how they integrate with skin tissue, absorption time, and possible side effects. This provides more comprehensive theoretical support for minimally invasive cosmetic treatments. By simulating the dynamic changes of materials, it can predict potential adverse reactions or unsatisfactory results in advance, reducing surgical risks and increasing patient confidence in the treatment's effectiveness.

[0115] In a preferred embodiment of the present invention, step 14 further includes:

[0116] Step 143: Based on the recognition results of the facial micro-plastic surgery interaction effect model, and combined with the characteristic data of micro-plastic surgery filling or dissolving materials, a simulation is performed; based on the simulated effect of the filling or dissolving material interaction effect on the patient's face, a postoperative effect visualization image is generated to show the improvement effect of the filling or dissolving materials on facial defect features; based on the displayed effect, feedback is provided in real time and the plan is adjusted in a timely manner to generate a personalized facial micro-plastic surgery final plan.

[0117] In this invention, simulations are performed based on the recognition results of a micro-plastic surgery interaction effect model. First, a three-dimensional image of the patient's face and historical data are input. Using a previously established facial micro-plastic surgery interaction effect model, the model identifies the patient's facial defects and the required filler or dissolving materials. Next, the model combines the characteristic data of the micro-plastic surgery filler or dissolving materials (such as material density, elasticity, absorption rate, and skin integration) to simulate the interaction effects of these materials on facial tissue. By simulating the interaction between materials and tissue, the system can predict the specific effects of different materials on the patient's face. For example, the simulation can show how filler materials change facial contours, how they improve wrinkles or depressions, or how dissolving materials reduce facial swelling or improve asymmetry.

[0118] Through in-depth simulation of material properties and interactive effects, doctors can gain a precise understanding of the effects of each material, avoiding poor results caused by material mismatch in actual treatment. Simulation based on the patient's specific facial condition allows for the customization of suitable minimally invasive cosmetic procedures for each patient, reducing unnecessary adjustments and incorrect treatment choices. By simulating the effects of filling or dissolving materials on the face, doctors can quickly select the most suitable materials and dosages, improving treatment outcomes.

[0119] After the simulation is complete, the system will generate a post-operative visualization image based on the material's interactive effects. This image presents the patient's expected facial results after treatment, demonstrating how the filler or dissolving material improves facial imperfections, such as smoothing facial contours, restoring firmness, or improving wrinkles and depressions. These effects are presented through 3D reconstruction, intuitively reflecting the material's actual effect on the face. The post-operative visualization image provides patients with an intuitive preview of the results, helping them better understand the expected changes after treatment, thereby increasing their confidence in the treatment. Patients and doctors can communicate better through the visualization image, ensuring that patients have clear expectations of the post-operative results and avoiding misunderstandings or dissatisfaction with the outcome. Through the simulated post-operative results, doctors can assess the degree of match between the material's actual effect and the patient's expectations, providing important reference for subsequent treatment decisions.

[0120] After the visualization is generated, patients or doctors can provide real-time feedback based on the displayed results. If a patient is dissatisfied with the simulated post-operative outcome, or if a doctor believes that the effect of certain materials or filled areas is unsatisfactory, the treatment plan can be adjusted promptly based on the feedback. This means that the type of material, the amount of filler, or the treatment area can be adjusted to achieve a more ideal result;

[0121] When the filling effect in a certain area is too full or has an unnatural appearance, the doctor can adjust the amount of filling material or choose different materials for further optimization; the real-time feedback and adjustment function can optimize according to the patient's actual needs, ensuring that the treatment plan is more personalized and precise, thereby improving patient satisfaction.

[0122] By simulating and adjusting the treatment plan in a timely manner, the need for postoperative repair can be reduced because the plan has already been optimized and adjusted multiple times based on the results before the actual operation. Each patient has different facial features and needs, and real-time adjustments allow the treatment plan to flexibly adapt to the changes of different patients, ensuring personalized treatment results.

[0123] After simulation, visualization, and adjustments, the final treatment plan was confirmed and generated. This is a personalized facial minimally invasive cosmetic surgery plan that has undergone multiple optimizations. The plan considers the patient's facial features, defects, required filler or dissolving materials, and optimal treatment results to ensure ideal cosmetic outcomes. Through preceding simulations and feedback adjustments, the final plan better aligns with the patient's needs and the doctor's professional advice, guaranteeing both treatment effectiveness and patient safety. Personalized minimally invasive cosmetic surgery plans maximize the fulfillment of patients' aesthetic needs and expectations, avoiding inconsistencies or dissatisfaction that may result from standardized treatment plans. Based on extensive simulation data, feedback, and adjustments, the final plan boasts higher precision, ensuring the minimally invasive surgery achieves optimal results.

[0124] In a preferred embodiment of the present invention, step 143 further includes:

[0125] Step 1431: Based on the characteristic data of different filling or dissolving materials, obtain the deformation effect of the filling or dissolving materials by using σ=E·ε; where σ represents stress, E represents the elastic modulus of the material, and ε represents strain.

[0126] Step 1432, based on the deformation effect of the filling or dissolving material, through The diffusion degree of the filling or dissolving material in the subcutaneous tissue gap is obtained to describe the adaptation and distribution of the material on the face; and combined with the facial micro-plastic surgery interactive effect model, the effect of the material filling or dissolving on the facial surface morphology is obtained.

[0127] In the formula, δ represents the diffusion degree of the filling or dissolving material in the subcutaneous tissue interstitial space; F represents the deformation force generated by the filling or dissolving material; k skin This represents the skin's elasticity coefficient.

[0128] In this invention, step 1431 involves analyzing the characteristic data of different filling or dissolving materials, focusing on the deformation effect of the materials. Parameters such as stress (σ), elastic modulus (E), and strain (ε) are used to simulate the material's behavior on the face based on these physical quantities. Stress describes the influence of external forces on the material, the elastic modulus represents the material's elastic properties, i.e., its ability to deform under external forces, while strain indicates the degree of deformation of the material under stress.

[0129] This analysis allows for the prediction of the deformation effects of each filler or dissolving material when applied to the face. For example, it reveals how the material adjusts its shape in response to facial muscle activity or how it produces different effects in different areas. This simulation reflects the material's adaptability to facial skin and its shaping effect on facial contours. By modeling the material's physical properties, deformation in different facial areas can be accurately predicted, helping doctors understand the actual effects after application. Calculating stress, elastic modulus, and strain effectively prevents unnatural facial changes, such as asymmetrical or abrupt filling effects. Based on the deformation characteristics of different materials, doctors can select the most suitable material for each patient to ensure optimal treatment results.

[0130] Step 1432, based on the deformation effect of the filling or dissolving material in Step 1431, further calculates the diffusion degree of the material in the subcutaneous tissue space. Diffusion degree reflects the distribution characteristics of the material in the subcutaneous tissue space, mainly describing how the material expands, adapts, and distributes in different areas within the subcutaneous tissue space. This process considers factors such as skin elasticity, texture, and thickness to evaluate the distribution of the material on the face and their interactions.

[0131] Building upon previous interactive models of facial micro-plastic surgery, the system predicts the specific impact of filling or dissolving materials on facial surface morphology based on the material's diffusion rate. For example, filler materials may create noticeable bumps in certain areas, while dissolving materials may adjust facial smoothness through slow dispersion. Through this comprehensive analysis, the system can provide each patient with a more precise facial contour improvement plan.

[0132] Diffusion calculations help doctors better understand how materials are evenly distributed in the subcutaneous tissue spaces of the face, avoiding uneven results and ensuring a balanced and natural facial appearance. By analyzing the diffusion and distribution of filler or dissolving materials on the face, doctors can more precisely control the use of each material, ensuring personalized and efficient treatment plans. Combining diffusion and micro-plastic surgery interaction models, it is possible to predict the impact of each material on facial appearance, thereby providing patients with the most suitable treatment plan.

[0133] The combined use of steps 1431 and 1432 provides a detailed prediction of the material's effect on the face by analyzing the deformation and diffusion of the filling or dissolving material. First, the material's deformation is assessed using a physical model to ensure that the material's selection and application naturally adapt to the facial structure. Then, the diffusion is calculated to optimize the material's distribution on the face, thereby predicting the specific impact of the treatment on facial morphology. The core purpose of this process is to accurately simulate the material's actual effect before treatment, ensuring that each patient's personalized treatment plan is precise and effective, and avoiding unnatural or asymmetrical problems that may occur during surgery.

[0134] In a preferred embodiment of the present invention, step 143 further includes:

[0135] Step 1433: Based on the influence of the facial surface morphology after filling or dissolving, use B-spline deformation to smooth the facial lines of the 3D facial micro-plastic surgery image generated by the generative AI model.

[0136] Step 1434, through I = I t +I d +I s To obtain a realistic representation of the post-operative lighting effect on the face and to map facial textures to improve the realism of the visualization; the facial textures include skin tone uniformity, texture coherence, and smoothness; where, I t I d and I s These represent the brightness of ambient light, diffuse light, and specular light, respectively, while I represents the postoperative illumination of the face as actually displayed.

[0137] Step 1435: Based on the visualized image, view the effect of the micro-plastic surgery to adjust the parameters of material filling amount, position, and dissolution effect; and update the generative AI model to generate the final visualized image of the postoperative effect of facial micro-plastic surgery.

[0138] In this invention, step 1433, based on the facial surface morphology influence results generated in step 1432, uses B-spline deformation technology to smooth the facial 3D image generated by the generative AI model. B-spline deformation effectively smooths facial lines, optimizes facial contours, and ensures a softer, more natural result after minimally invasive cosmetic procedures. Through B-spline deformation, facial lines and contours become smoother and more natural, avoiding the stiff and unnatural effects caused by minimally invasive procedures. This process ensures that the facial minimally invasive cosmetic results are more ergonomic and visually more comfortable.

[0139] Step 1434 aims to add realistic lighting effects to the generated 3D facial image, simulating how light affects the face in the actual post-operative environment. Furthermore, facial texture is mapped, including skin tone uniformity, texture coherence, and smoothness, to enhance the image's realism. This is achieved by considering ambient light (IA). t ), diffuse reflection light (I) d ) and specular reflection (I s The brightness of the image allows for a more accurate representation of the face's visual effects under different lighting conditions; the mapping of lighting effects and facial textures greatly enhances the realism and visualization of the image, making the virtual image closer to the post-operative result. This process helps doctors and patients understand the effects of minimally invasive cosmetic procedures more intuitively, ensuring that the expected post-operative results are close to reality.

[0140] Step 1435 allows users to view and adjust the effects of the minimally invasive cosmetic procedure through a visualized image, including parameters such as filler volume, placement, and dissolution effect. This step provides an interactive platform, allowing users to adjust the image in real time until the ideal result is found. Simultaneously, the system feeds back and updates the generative AI model, generating the final post-operative image of the facial minimally invasive cosmetic procedure. The advantage of this step lies in its interactivity and real-time nature, making the minimally invasive cosmetic procedure more personalized and flexible. Through continuous adjustments and feedback, users can precisely control the effects of the minimally invasive procedure, avoiding unsatisfactory results due to errors or imperfect predictions. Ultimately, the generated visualized image accurately displays the post-operative effect, helping patients make more informed decisions.

[0141] like Figure 2 As shown, a facial minimally invasive cosmetic surgery effect simulation system based on generative AI includes:

[0142] Data acquisition module: It is used to acquire various three-dimensional facial image data of patients to obtain a three-dimensional facial image dataset; extract and label the facial defect features of each data item in the three-dimensional facial image dataset to obtain a defect labeling dataset, wherein the defect labeling data represents the marking of defect feature data existing on the patient's face;

[0143] Data recognition module: It is used to identify the defect feature data of the patient's facial three-dimensional image dataset and the defect labeling dataset collected in history, and to simulate the initial micro-plastic surgery plan based on the patient's facial defect feature data through a generative AI model;

[0144] Real-time adjustment module: It is used to allow patients to intuitively feel the comparison effect of facial improvement before and after surgery based on the initial micro-plastic surgery plan; based on the comparison results and real-time interactive feedback with the patient, it adjusts the surgical details in multiple rounds according to the patient's expected effect, so as to set the real-time adjustment strategy and generate the final facial micro-plastic surgery plan;

[0145] Effect Display Module: Based on the final facial micro-plastic surgery plan, it constructs the facial micro-plastic surgery interactive effect model to obtain the interaction effect between the facial micro-plastic surgery filling or dissolving materials and the human face, and updates the generative AI model to obtain the final visual display image of the face after surgery.

[0146] When the functions of the above-mentioned modules are implemented as software functional units and used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, external hard drives, ROM, RAM, magnetic disks, or optical disks.

[0147] Therefore, the object of the present invention can also be achieved by running a program or a set of programs on any computing device. The computing device can be a known general-purpose intelligent device. Therefore, the object of the present invention can also be achieved simply by providing a program product containing program code implementing the method or system. That is, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any known storage medium or any storage medium developed in the future. It should also be noted that in the apparatus and method of the present invention, it is obvious that the steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent to the present invention. Furthermore, the steps performing the above series of processes can naturally be performed in chronological order as described, but are not necessarily required to be performed in chronological order. Some steps can be performed in parallel or independently of each other.

Claims

1.A method of simulating a facial microplastic surgery effect based on generative AI, characterized by, The method comprises the following steps: Collecting a plurality of patient facial three-dimensional image data to obtain a facial three-dimensional image data set; extracting and labeling the facial defect features of each data item in the facial three-dimensional image data set to obtain a defect labeling data set, wherein the defect labeling data represents the labeling of the defect feature data present on the patient's face; wherein the extraction and labeling comprises: The three-dimensional image data in the three-dimensional image data set is standardized to a unified range through centering and scaling processing to obtain a normalized three-dimensional image data set; the three-dimensional image models in the normalized three-dimensional image data set are aligned through an iterative closest point algorithm to make different facial images in the same reference coordinate system; the three-dimensional image data items in the normalized three-dimensional image data set are image-labeled three-dimensional positioning key points through statistical analysis of the facial three-dimensional shape; a plurality of key points are arranged around the facial defect area; and the facial defect area features are extracted through three-dimensional point cloud technology to obtain the facial defect area features; Based on the facial defect area features extracted from the three-dimensional image, a sliding window algorithm is used to find the defect area and perform labeling processing to obtain defect labeling data; a plurality of defect labeling data are obtained based on a plurality of different facial three-dimensional images to form a defect labeling data set; based on the patient facial three-dimensional image data set and the defect labeling data set collected historically, a generative AI model is used to identify the defect feature data present on the patient's face and simulate an initial microplasty scheme for the patient's facial defect feature data; Based on the initial microplasty scheme, the patient can intuitively feel the contrast effect of facial improvement before and after the operation; based on the comparison result and real-time interaction feedback with the patient, the operation details are adjusted according to the patient's expected effect to set a real-time adjustment strategy to generate a final facial microplasty scheme; the real-time adjustment strategy comprises: Based on the initial microplasty scheme of the patient's facial defect feature data, the patient's facial defect degree is obtained; A pre-set facial flatness threshold range is compared with the patient's facial defect degree, and based on the comparison result, the patient's facial defect degree is obtained as concave or convex; By , to obtain the convexity and concavity degree of the facial defect, wherein R1 and R2 respectively represent the principal curvature radius of the convex and concave area; when the facial concave feature appears, the facial concave area is filled by a filling material, so that the concave area reaches a preset facial flatness threshold range; when the facial convex feature appears, the facial convex area is dissolved by a dissolving material, so that the convex area reaches the preset facial flatness threshold range, to obtain a micro-liposuction secondary scheme. Based on the final facial microplasty scheme, a facial microplasty interaction effect model is constructed to obtain the interaction effect of the facial microplasty filling or dissolving material with the human face, and the generative AI model is updated to obtain a final postoperative visual display image of the face: Obtain the microplasty filling or dissolving material characteristic data that fits the human tissue to obtain a material characteristic data set; construct a training set with the historical facial three-dimensional image data set and the material characteristic data set to generate a facial microplasty interaction effect model; the facial microplasty interaction effect model outputs an interaction effect recognition result of the facial image required microplasty filling or dissolving material; Input the newly acquired patient facial three-dimensional image into the facial micro-orthopedic interaction effect model to output facial defect feature and required filling or dissolving material interaction effect prediction results of the newly acquired patient facial three-dimensional image; based on the interaction effect prediction results, obtain simulation material and facial tissue interaction data through dynamic change characteristics of the filling or dissolving material in the patient body. 2.The method of claim 1, wherein, Scan and collect patient facial image data, and perform three-dimensional modeling on the patient facial image to generate facial three-dimensional image data; and integrate multiple facial three-dimensional image data into the facial three-dimensional image data set; And retain the two-dimensional image data of the upper, lower, left and right directions of each facial three-dimensional image to obtain the micro-orthopedic reference image data set. 3.The method of claim 2, wherein the method further comprises: Based on the micro-orthopedic secondary scheme, according to the patient's feedback on the satisfaction of the operation effect, the description of the facial change requirement and the personal preference, based on the generative AI model, the adjustment simulation result is output; The adjustment simulation result includes adjusting the facial contour, skin tightness, fat distribution and filling depth to dynamically optimize the adjustment of the patient's facial coordination effect in multiple dimensions; based on the adjustment simulation result, the final personalized facial micro-orthopedic scheme is recommended, and the iterative optimization is carried out combined with the individual difference of the patient to obtain the micro-orthopedic tertiary scheme. 4.The method of claim 1, wherein the method further comprises: receiving a user input of a target face image; and generating a target face image based on the user input. Based on the identification result of the facial micro-orthopedic interaction effect model, and combined with the simulation of the micro-orthopedic filling or dissolving material characteristic data; according to the simulation effect of the filling or dissolving material interaction effect on the patient's face, the postoperative effect visualization image is generated to show the improvement effect of the filling or dissolving material on the facial defect feature; based on the display effect, the scheme is adjusted in real time to generate the final personalized facial micro-orthopedic scheme. 5.The method of claim 4, wherein the method is based on generative AI. According to the characteristic data of the different filling or dissolving materials, by obtaining the deformation effect of the filling or dissolving materials; in the formula, σ represents stress, E represents the characteristic data of the elastic modulus of the material, and ε represents strain. based on the deformation effect of the filling or dissolving material, by obtaining the diffusion degree of the filling or dissolving material in the interstitial space of the subcutaneous tissue to describe the adaptation and distribution of the material on the face; and combining the face micro-integer interaction effect model to obtain the influence result of the material filling or dissolving on the surface morphology of the face; In the formula, represents the diffusion degree of the filling or dissolving material in the interstitial space of the subcutaneous tissue; F represents the deformation force generated by the filling or dissolving material; k skin represents the skin elasticity coefficient. 6.The method of claim 5, wherein the method is based on generative AI. According to the influence result of the filling or dissolving facial surface morphology, the B-spline deformation is used to smooth the facial line of the facial micro-orthopedic three-dimensional image generated by the generative AI model; By , to obtain the postoperative lighting effect of facial real demonstration, and map the facial texture to improve the realism of visualization; the facial texture includes skin color uniformity, texture contact degree and flatness; in the formula, I t , I d and I s respectively represent the brightness of ambient light, diffuse reflection light and specular reflection light, and I represents the postoperative lighting brightness of facial real demonstration; Based on the visualization image, the micro-orthopedic effect is viewed to adjust the filling amount, position, dissolving effect and other parameters of the material; And feedback to update the generative AI model to generate the final facial micro-orthopedic postoperative effect visualization image. 7.A facial microplastic surgery effect simulation system based on generative AI, characterized by, The system sets an electronic device including a memory, a processor and a generative AI-based facial micro-orthopedic surgery effect simulation method program stored on the memory and executable on the processor, and the generative AI-based facial micro-orthopedic surgery effect simulation method program is executed by the processor to realize the steps of the generative AI-based facial micro-orthopedic surgery effect simulation method of any one of claims 1-6, and the system comprises: A data acquisition module is used to acquire multiple patient facial three-dimensional image data to obtain a facial three-dimensional image data set; the facial defect features of each data item in the facial three-dimensional image data set are extracted and labeled to obtain a defect labeling data set, and the defect labeling data represents the labeling of the defect feature data existing in the patient's face; The data recognition module is configured to recognize defect feature data existing in the three-dimensional image of the patient's face based on the historical collected three-dimensional image data set of the patient's face and the defect calibration data set, and simulate an initial microplasty scheme for the patient's face defect feature data through the generative AI model. The real-time adjustment module is configured to enable the patient to intuitively feel the contrast effect of the improvement of the face before and after the operation based on the initial microplasty scheme, interact with the patient in real time based on the comparison result, and adjust the operation details according to the expected effect of the patient in multiple rounds to set a real-time adjustment strategy and generate a final microplasty scheme. The effect display module is configured to obtain the interactive effect of the filling or dissolving material of the microplasty and the human face by constructing the microplasty interactive effect model based on the final microplasty scheme, and update the generative AI model to obtain the final visual display image of the face after the operation.

Images