Intelligent analogue simulation method and system for facial micro plastic surgery operation

Abstract

The invention discloses an intelligent analogue simulation method and system for facial micro plastic surgery operation, and relates to the technical field of image processing. According to the intelligent analogue simulation method, high-precision facial data are obtained, and facial defect parts are accurately calibrated; after normalization processing is carried out on face three-dimensional image data, key face features are extracted through a deep learning algorithm, and a basis is provided for a micro-surgery scheme. Different types of data are integrated through an intelligent fusion technology to form a comprehensive facial analysis atlas, and virtual simulation and operation effect evaluation are supported. Then, a personalized operation scheme is automatically generated, the postoperative effect is predicted, accurate operation strategies such as material selection, injection depth and complication risk are provided, and the operation risk is remarkably reduced. The system adjusts the scheme in real time in combination with patient feedback, and performs postoperative effect evaluation to ensure that the effect accords with expectation.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly to an intelligent simulation method and system for facial microsurgery operations. Background Art

[0002] In recent years, with the development of intelligent technology and computer simulation technology, more and more medical fields have begun to use virtual simulation technology for preoperative simulation and planning. Especially in the field of facial microsurgery, intelligent simulation based on the three-dimensional facial data of patients has gradually attracted attention;

[0003] However, in existing simulation methods, due to the large differences in facial muscles, fat, skin and other tissues among different individuals, the accuracy of the simulation effect of facial microsurgery is limited, resulting in difficulties in avoiding the risk of nerve damage caused by injection operations during facial microsurgery simulation and accurately grasping the content of filling or dissolving materials, which affects the acceptance degree of patients for the simulation results of microsurgery;

[0004] In summary, how to incorporate more detailed facial anatomical information into the intelligent simulation process, achieve accurate positioning of facial defect sites, and reduce the risks of facial nerve damage and material injection is an urgent problem to be solved and optimized for the intelligent simulation system. Summary of the Invention

[0005] The present invention provides an intelligent simulation method and system for facial microsurgery operations, which solves the technical problem of how to incorporate more detailed facial anatomical information into the intelligent simulation process, achieve accurate positioning of facial defect sites, and reduce the risks of facial nerve damage and material injection.

[0006] To solve the above technical problems, the present invention provides an intelligent simulation method and system for facial microsurgery operations, and the specific technical solutions are as follows:

[0007] In a first aspect, an intelligent simulation method for facial microsurgery operations includes the following steps:

[0008] Obtain the facial data of a patient, various facial defect feature data, and various defect structure anatomical information data; and perform normalization, feature extraction, and encoding processing on each data;

[0009] Fuse the processed data and perform virtual simulation to obtain various microsurgery plan information data; construct a facial microsurgery prediction model with the facial data, various facial defect feature data, various defect structure anatomical information data, and various microsurgery plan information data;

[0010] The facial microplastic prediction model can output the prediction result of the microplastic scheme information data corresponding to the defect structure anatomical information data of the facial defect feature data;

[0011] Based on the prediction result of the facial microplastic prediction model, a personalized facial microplastic prediction scheme is obtained; based on the facial microplastic prediction scheme, a precise surgical operation strategy is obtained for doctor-patient interaction feedback and evaluation of the microplastic postoperative effect.

[0012] As a further optimized scheme of the present invention, patient facial data is acquired to construct a three-dimensional facial model of the patient; the three-dimensional facial model is analyzed to obtain various facial defect feature data, which are assembled into a facial defect feature data set;

[0013] The various facial defect feature data are located and marked to obtain a defect marked data set; each facial defect feature data item in the defect marked data set is anatomically structured to obtain a defect structure anatomical information data set.

[0014] As a further optimized scheme of the present invention, the three-dimensional facial model is analyzed to obtain various facial defect feature data assembled into a facial defect feature data set, including:

[0015] Based on the patient facial data to obtain three-dimensional facial information data of the patient; the three-dimensional facial information data is preprocessed to obtain preprocessed three-dimensional facial information;

[0016] The preprocessed three-dimensional facial information is subjected to point cloud or mesh processing to construct a three-dimensional facial model; facial feature extraction is performed on the three-dimensional facial model to obtain facial feature data;

[0017] Facial defect detection is performed on the facial feature data to obtain facial defect feature data; different types of facial defect feature data are summarized to obtain a facial defect feature data set.

[0018] As a further optimized scheme of the present invention, each facial defect feature data item in the defect marked data set is anatomically structured to obtain a defect structure anatomical information data set, including:

[0019] The defect structure anatomical information data set includes the spatial position and structural information data of the facial defect skin layer, subcutaneous tissue layer, muscle layer, nerve direction, and blood vessel distribution; the spatial position and structural information data includes the adhesion relationship and layer thickness width between the spatial positioning and structural information of each layer;

[0020] According to the obtained spatial position and structural information data, combined with the patient's skin laxity, fat distribution and fat content, perform simulation; so that the facial simulated three-dimensional image matches the facial anatomical structure information data to obtain a matching result;

[0021] Based on the matching result, obtain the defective structure anatomical image data; divide the anatomical structure of the defective area of each defective structure anatomical image data, and partition and mark the divided defective area to accurately position the anatomical data items of each defective area.

[0022] As a further optimization scheme of the present invention, dividing the anatomical structure of the defective area of each defective structure anatomical image data, and partitioning and marking the divided defective area to accurately position the anatomical data items of each defective area, including:

[0023] For the defective area in each defective structure anatomical image data, through C i ={p|I(p)=1 + C(p, C j )}, connect adjacent pixel points together to form different defective partitions;

[0024] wherein, C i represents the i-th connected region, p represents a pixel point in the image, I(p) represents the pixel vector of the pixel, and C(p, C j ) represents the connection vector between pixel p and region C j ;

[0025] Perform topological analysis on different defective partitions according to the spatial position and structural information data of the facial defective skin layer, subcutaneous tissue layer, muscle layer, nerve orientation and blood vessel distribution, so that the defective area is segmented into different structural layers, and remove the defective areas with noise and broken connections;

[0026] Perform hierarchical marking on different structural layers to accurately obtain the intertwined points of the muscle layer, nerve orientation and blood vessel distribution in the same defective area.

[0027] As a further optimization scheme of the present invention, perform shape detection on the obtained facial defects to obtain two defective feature shapes of depression and protrusion, and respectively pass through and to obtain the spatial volume of the facial defect; where V1 represents the spatial volume of the depression feature, V2 represents the spatial volume of the protrusion feature, h1 represents the depth of the depression, h2 represents the height of the protrusion defect feature, r1 represents the inner diameter of the depression; r2 represents the bottom inner diameter of the protrusion; L represents the top width of the protrusion;

[0028] According to the spatial volume of the facial defect, combined with the patient's skin laxity, fat distribution and fat content, through G=(V×ρ 材料) × (1 + k) to obtain the content of the injectable minimally invasive plastic material required for facial defects; where G represents the value of the content of the injectable minimally invasive plastic material required, V represents the volume V1 of the sunken feature space or the volume V2 of the protruding feature space, and ρ 材料 represents the density of the injectable material, and k represents the correction coefficient of the facial absorption rate;

[0029] According to the intersection points of the muscle layer, nerve direction, and blood vessel distribution in the same defect area, inject the minimally invasive plastic material precisely into the facial defect area of the patient. Through the display of an animated simulation, the patient can intuitively receive the surgical process and interact with the doctor for feedback to avoid surgical risks.

[0030] As a further optimized solution of the present invention, the facial minimally invasive plastic prediction model includes:

[0031] Generate structure data from the obtained multiple facial defect feature data, multiple defect structure anatomical information data, and minimally invasive plastic plan information data sets, and encode the structure data into sequence data to train the facial minimally invasive plastic prediction model;

[0032] Input the sequence data into the facial minimally invasive plastic prediction model; the facial minimally invasive plastic prediction model includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer. Transmit the intermediate representation data of multiple hidden layers to the output layer, and the output layer outputs the recognition result of the minimally invasive plastic plan information data corresponding to the defect structure anatomical information data representing the facial defect feature data set;

[0033] Input at least one newly obtained facial defect feature data into the facial minimally invasive plastic prediction model to output the prediction result of the minimally invasive plastic plan information data corresponding to the defect structure anatomical information data representing the facial defect feature data.

[0034] As a further optimized solution of the present invention, according to the prediction result of the facial minimally invasive plastic prediction model, obtain the corresponding minimally invasive plastic plan information data, and compare the corresponding minimally invasive plastic plan information data with the minimally invasive plastic plan information data of historical similar facial defect features to obtain a comparison result;

[0035] Based on the comparison result, obtain the minor difference data, perform a correlation analysis on the minor difference data to obtain the minor difference influencing factor; further iteratively update the parameters of the facial minimally invasive plastic prediction model according to the minor difference influencing factor to accurately obtain the prediction result of the personalized surgical plan for minimally invasive plastic surgery;

[0036] The minor difference influencing factor includes the patient's face shape features, facial features, age, gender, and personalized needs, and adjusts the prediction result of the output minimally invasive plastic plan information data in real time to obtain a preliminary surgical plan;

[0037] Generate feedback information of surgical effect pictures and facial change curves according to the prediction results of the micro - plastic surgery plan information data, so as to iteratively optimize the preliminary surgical plan and obtain the final micro - plastic surgery plan.

[0038] As a further optimization scheme of the present invention, the surgical precise operation strategy includes:

[0039] Customize the micro - plastic surgery path according to the differences in the interweaving points of the muscle layer, nerve direction and blood vessel distribution of the facial defect anatomical structure in the same defect area, as well as the directions of the three major facial nerves, so as to avoid the densely intertwined nerve areas.

[0040] In a second aspect, the system is provided with an electronic device including a memory, a processor, and a program of an intelligent simulation method for facial micro - plastic surgery operation stored on the memory and executable on the processor. When the program of the intelligent simulation method for facial micro - plastic surgery operation is executed by the processor, it realizes the steps of an intelligent simulation method for facial micro - plastic surgery operation. The system includes:

[0041] Data acquisition module: It is used to acquire patient facial data, various facial defect feature data, and various defect structure anatomical information data;

[0042] Pre - processing module: It is used to perform normalization, feature extraction, and encoding processing on each data;

[0043] Simulation module: It is used to fuse the processed data and perform virtual simulation to obtain various micro - plastic surgery plan information data;

[0044] Model construction module: It is used to construct a facial micro - plastic prediction model from facial data, various facial defect feature data, various defect structure anatomical information data, and various micro - plastic surgery plan information data;

[0045] Scheme prediction module: It is used for the facial micro - plastic prediction model to output the prediction results of the micro - plastic surgery plan information data corresponding to the defect structure anatomical information data of the facial defect feature data;

[0046] Feedback evaluation module: It is used to obtain a personalized facial micro - plastic prediction plan based on the prediction results of the facial micro - plastic prediction model; and obtain a surgical precise operation strategy based on the facial micro - plastic prediction plan for doctor - patient interaction feedback and evaluation of the micro - plastic surgery effect.

[0047] The present invention has at least the following beneficial effects: The present invention obtains high-precision data of the patient's face through a variety of advanced image acquisition technologies (such as 3D facial scanning, CT or MRI imaging, etc.), including feature information such as the anatomical structure, fat layer, muscle layer, nerve distribution, and skin laxity of the face; Facial defect information (such as wrinkles, facial depressions, laxity, etc.) and the specific location and depth of the defects are accurately calibrated and matched with the patient's anatomical feature data to ensure that each patient's personalized defect site can be clearly defined.

[0048] By normalizing facial data from different sources and in different formats, the data has a unified standard, thus avoiding the deviation caused by inconsistent data sources. This process can effectively improve the accuracy of data processing and the reliability of subsequent modeling; Through deep learning algorithms or machine learning models, facial features are extracted from the facial data, and key features of the face (such as fat layer thickness, skin laxity, muscle position, etc.) are extracted. These features can provide important bases for the generation of subsequent minimally invasive plastic surgery plans.

[0049] Different types of data (such as facial anatomical structure data, defect feature data, anatomical information data, etc.) are integrated through intelligent data fusion technologies (such as multi-modal learning) to form an all-round facial analysis atlas. This process can ensure that the relationships between data are fully explored and provide accurate bases for generating surgical plans; After virtual simulation of the processed data, the effects of minimally invasive plastic surgery on the face can be simulated, and the effects can be evaluated through different operation methods and surgical plans in the virtual environment. This process can realize the comparison and optimization of multiple minimally invasive plastic surgery plans and help doctors select the best plan; According to the patient's facial features and defect data, the system can automatically generate multiple minimally invasive plastic surgery plans and provide personalized treatment suggestions based on the expected effects and risks of each plan. This can greatly reduce the error of preoperative judgment and reduce unnecessary surgical risks.

[0050] Based on big data and deep learning models, all processed data (facial anatomical data, defect features, anatomical information, surgical plans, etc.) are fused and a facial minimally invasive plastic surgery prediction model is constructed. This model can predict the minimally invasive plastic surgery plan for the facial defect site according to the specific situation of the patient and output the corresponding surgical effect; Through the prediction results of this model, doctors can obtain detailed personalized surgical plans, including the usage amount of materials, injection depth, precise injection point location, possible complications, etc. This prediction process can effectively avoid the uncertainty in traditional surgical plans and greatly reduce the surgical risks.

[0051] Based on the output results of the facial micro - plastic prediction model, personalized micro - plastic surgery plans can be generated. Each plan is tailored according to factors such as the patient's facial features, defective parts, and muscle and fat distribution to ensure maximum compliance with the patient's needs; generate precise surgical operation strategies based on the personalized micro - plastic plan, which includes specific surgical steps, material selection, injection positions, injection depths, and pressures, etc. This operation strategy can provide an accurate execution framework for doctors and reduce randomness and risks during the operation.

[0052] Through the interaction between the system and the patient, the patient can provide real - time feedback on their needs and expectations, and the doctor can adjust the surgical plan according to the feedback information. In addition, the real - time evaluation of the postoperative effect, combined with virtual simulation technology, can help doctors compare and adjust the postoperative effect to ensure that the postoperative effect meets expectations. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a schematic flow chart of an intelligent simulation method for facial micro - plastic surgery operations provided by an embodiment of the present invention;

[0054] Figure 2 is a schematic diagram of an intelligent simulation system for facial micro - plastic surgery operations provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0055] The following further describes the present application in detail with reference to the drawings. It is necessary to point out here that the following detailed embodiments are only used to further illustrate the present application and should not be construed as limiting the protection scope 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.

[0056] An intelligent simulation method and system for facial micro - plastic surgery operations provided by this embodiment are specifically implemented as follows:

[0057] As Figure 1 shown, an intelligent simulation method for facial micro - plastic surgery operations includes the following steps:

[0058] Step 11, obtain the patient's facial data, various facial defect feature data, and various defective structure anatomical information data;

[0059] Step 12, perform normalization, feature extraction, and encoding processing on each data;

[0060] Step 13, fuse the processed data and perform virtual simulation to obtain various micro - plastic surgery plan information data;

[0061] Step 14: Construct a facial microplastic surgery prediction model with the facial data, various facial defect feature data, various defect structure anatomical information data, and various microplastic surgery plan information data.

[0062] Step 15: The facial microplastic surgery prediction model can output a prediction result of the microplastic surgery plan information data corresponding to the defect structure anatomical information data of the facial defect feature data.

[0063] Step 16: Based on the prediction result of the facial microplastic surgery prediction model, obtain a personalized facial microplastic surgery prediction plan; based on the facial microplastic surgery prediction plan, obtain a precise surgical operation strategy for doctor-patient interaction feedback and evaluation of the effect after microplastic surgery.

[0064] In the implementation of the present invention, in Step 11, the facial data of the patient is obtained through a facial scanning device, digital facial photography, 3D modeling, or other medical imaging technologies. These data may include the patient's facial contour, skin condition, and facial proportions; based on the facial scanning data, the features of facial defects are extracted, such as wrinkles, scars, uneven skin color, and facial symmetry; detailed anatomical data of the facial structure are obtained through medical imaging or professional equipment, including but not limited to bone structure, muscle distribution, and skin layer information data; the accurate data of the patient's face provides a real basis for formulating subsequent microplastic surgery plans, ensuring the personalization of diagnosis and treatment; by comprehensively collecting facial data at different levels, facial problems of the patient can be accurately analyzed to avoid missed diagnoses; providing detailed facial anatomical structure data for doctors helps to better understand the root causes of facial defects.

[0065] In Step 12, different types of data (facial images, defect features, anatomical information) are normalized to be within a unified scale range, usually by adjusting the data to between 0 and 1, reducing the impact of different data scales on model training; through image processing, machine learning, or deep learning algorithms, key facial features and defect information are extracted, such as facial expression features, skin condition, and the specific type and location of defects; the extracted features are converted into a numerical format that can be processed by machines (such as numerical encoding, one-hot encoding, etc.), which helps subsequent algorithms process the data; normalization helps to unify the scales between different data types, making the machine learning model more efficient and stable during training; feature extraction can screen out the most distinguishable data points, improving the prediction accuracy of the model; after encoding processing, the data can be effectively read by the machine learning model, ensuring the smooth operation of subsequent algorithms.

[0066] Step 13 fuses various types of data (facial data, defect features, anatomical information) to generate a comprehensive dataset; these data can be combined through weighted average, principal component analysis (PCA), and multimodal learning methods to produce a more comprehensive patient facial information; using computer graphics and virtual reality technologies, surgical simulations are performed based on the fused data to display the possible impacts of different minimally invasive plastic surgery plans on the patient's face, generating multiple surgical plan information; data fusion can synthesize multi-dimensional information to provide doctors with a more accurate overall picture of facial features and defects; virtual simulation allows patients to see in advance the impacts of different plans on facial appearance, thus helping doctors provide personalized minimally invasive plastic surgery suggestions for patients; virtual simulation can predict surgical effects and help patients and doctors make more appropriate choices.

[0067] Step 14 inputs all the data (facial data, defect features, anatomical information) obtained in Steps 11 and 12 together with the surgical plan information generated by virtual simulation into a machine learning model. By training these data, a facial minimally invasive plastic surgery prediction model is established. This model will be able to predict suitable minimally invasive plastic surgery plans based on the input patient facial information; through the training and learning of the model, it can quickly and accurately analyze the patient's facial information and give personalized surgical suggestions; reduce the subjective errors of doctors in the process of data processing and plan selection, and improve the accuracy of medical decisions.

[0068] Step 15 The model makes predictions based on the patient's facial defect features and anatomical data and outputs minimally invasive plastic surgery plans that match the patient's facial problems. These plans will include specific surgical types, treatment methods, and surgical steps; through the model output, doctors can customize the best treatment plan for each patient, avoiding one-size-fits-all treatment methods; accurate predictions can help select the most suitable minimally invasive plastic surgery plan for patients and improve surgical effects and patient satisfaction.

[0069] Step 16 generates a complete minimally invasive plastic surgery plan based on the prediction results of the facial minimally invasive plastic surgery prediction model and provides operation strategies for doctors; provides doctors with detailed operation guides on surgical steps, methods, and precautions. Patients can also participate in feedback and discussions through the doctor-patient interaction platform to help adjust the plan; based on the postoperative feedback information and evaluation results, further optimize future minimally invasive plastic surgery plans; through personalized minimally invasive plastic surgery prediction plans, it can ensure that each patient can obtain the best treatment effect; through postoperative evaluation and feedback, doctors can adjust the plan in real time and continuously improve the success rate of the surgery; through doctor-patient interaction, patients can more clearly understand the treatment process and effects, thereby enhancing trust and satisfaction.

[0070] In a preferred embodiment of the present invention, Step 11 further includes:

[0071] Step 111: Obtain the facial data of the patient and construct a three-dimensional facial model of the patient; analyze the three-dimensional facial model to obtain various facial defect feature data and compile them into a facial defect feature data set.

[0072] Step 112: Locate and mark the various facial defect feature data to obtain a defect marking data set; anatomize each facial defect feature data item in the defect marking data set to obtain a defect structure anatomy information data set.

[0073] In the implementation of the present invention, in Step 111, detailed data of the patient's face are obtained through various sensors (such as 3D scanners, stereo cameras, depth sensors, etc.). These data may include information such as facial contours, muscle structures, skin textures, etc.; using the obtained data, through computer vision and three-dimensional modeling techniques (such as point cloud reconstruction, surface fitting, etc.), the two-dimensional image data are converted into a three-dimensional facial model. This three-dimensional model can not only truly reflect the appearance of the face, but also present the depth information of the face, facilitating further analysis; through algorithms (such as image processing, machine learning, computer vision), the structure and features of the three-dimensional model are analyzed. Mainly extract feature points from the facial morphology (such as the contours of eyes, nose, mouth, the height of cheekbones, etc.), identify the subtle defects of the face, and construct a multi-dimensional defect feature data set; the three-dimensional facial model can help doctors or technicians more accurately identify the minor defects of the face and avoid the distortion that may occur in traditional two-dimensional images; the facial features of each person vary greatly, and three-dimensional modeling can provide detailed analysis for individuals, thus providing data support for subsequent treatment, plastic surgery, and customized medicine; through the facial defect feature data set obtained by analysis, defects can be automatically detected and marked, providing an effective basis for subsequent repair or treatment.

[0074] Based on the multi-dimensional defect data set obtained in Step 111, in Step 112, the system automatically marks the facial defect areas through algorithms, such as wrinkles, spots, moles, facial asymmetry, etc. These defect features are located in specific facial areas (such as forehead, corners of eyes, lips, etc.) to generate a defect marking data set; further detailed structural analysis of these defects is carried out. For example, analyze the depth of wrinkles, the size of moles, the degree of skin relaxation, etc., or obtain more detailed anatomical data through medical imaging techniques such as CT scans and MRIs. Convert these data into an "anatomy information data set", which is the detailed structural analysis data of facial defects; location marking can help doctors or technicians quickly and accurately identify and distinguish the specific locations and types of facial defects, thus contributing to the formulation of treatment plans; through the anatomical analysis of facial defects, more dimensional data can be obtained, helping to better understand the causes and nature of the defects; through detailed defect marking and anatomical information, a personalized treatment or repair plan can be formulated for the patient to ensure more accurate and efficient results.

[0075] In a preferred embodiment of the present invention, in step 111, the three-dimensional facial model is analyzed to obtain various facial defect feature data, which are compiled into a facial defect feature data set, including:

[0076] Step 1111, based on the patient's facial data to obtain the patient's three-dimensional facial information data; preprocess the three-dimensional facial information data to obtain the preprocessed three-dimensional facial information;

[0077] Step 1112, perform point cloud or mesh processing on the preprocessed three-dimensional facial information to construct a three-dimensional facial model; extract facial feature data from the three-dimensional facial model to obtain facial feature data;

[0078] Step 1113, perform facial defect detection on the facial feature data to obtain facial defect feature data; summarize different types of facial defect feature data to obtain a facial defect feature data set.

[0079] In the embodiment of the present invention, step 1111 obtains a high-precision three-dimensional facial model;

[0080] 3D scanning, using a three-dimensional scanner to scan the human face to obtain point cloud data;

[0081] Stereo vision, using two or more cameras to capture facial images from different angles and using a stereo matching algorithm to reconstruct a three-dimensional model;

[0082] Depth sensors, such as Kinect, LiDAR and other sensors, can also directly obtain the three-dimensional information of the face through depth maps; the generated three-dimensional facial information is usually in the form of point clouds or meshes.

[0083] Processing the original three-dimensional data includes:

[0084] Noise removal, noise or irregular points may appear during the three-dimensional model scanning process, which need to be removed by Gaussian filtering; point cloud registration, when obtaining point cloud data from multiple perspectives, align and merge them through the Iterative Closest Point (ICP) algorithm; mesh reconstruction, convert the point cloud into a triangular mesh model through Poisson reconstruction or Delaunay triangulation methods;

[0085] Based on the facial data preprocessed in step 1111, the geometric features of the face are crucial for identifying facial defects through step 1112; calibrate facial feature points (such as the positions of eyes, nose, mouth, etc.) on the three-dimensional facial model through the ASM algorithm or AAM algorithm; for two-dimensional images, through E = Σ i ||I(x i ) - T(x i )|| 2; Use the minimization objective function to find the key points; where I(x i ) represents the pixel value of the input image, T(x i ) indicates that the target model is in x i The texture of the position, E represents the key point of the facial features, and i represents the i-th facial feature point;

[0086] By calculating the geometric features of the facial mesh (such as curvature, normal vector, facial symmetry, etc.) to obtain deeper information, the Gaussian curvature and mean curvature of each geometric feature vertex of the facial network are used to describe the curvilinear shape of the face, and the normal vector of each vertex is used to describe the surface direction of the point;

[0087] Based on step 1112, step 1113 can be used to analyze the left-right symmetry of the face to detect deformities caused by diseases, trauma or congenital defects; symmetry detection can be evaluated by calculating the Euclidean distance between left-right symmetrical points; facial defects may include but are not limited to spots, scars, acne, depressions, deformities, etc. Detection of facial defects;

[0088] Use the Sobel operator or Laplacian operator to calculate the gradient of the image, detect edges or discontinuous areas, and use the local features of the depth map, texture image or mesh model to detect facial defects. For example, use morphological operations (dilation, erosion, etc.) to enhance the defective area, and then detect the defective area by thresholding;

[0089] The facial geometric features, symmetry detection, defect areas and other information extracted above are collected into a multi-dimensional feature data set. Each record of the data set may include:

[0090] Spatial coordinates of key points (x, y, z); facial geometric features (such as curvature, normal vector); facial symmetry information; defect markers (such as the presence or absence of scars and acne marks); detailed information on the size, shape, and location of defects; these feature data sets can be used for machine learning model training to help the system automatically identify and classify different types of facial defects.

[0091] In a preferred embodiment of the present invention, in step 112, each facial defect feature data item in the defect labeling data set is subjected to structural dissection to obtain a defect structural dissection information data set, including:

[0092] Step 1121, the defect structure anatomical information data set includes spatial position and structural information data of facial defect skin layer, subcutaneous tissue layer, muscle layer, nerve direction and blood vessel distribution; the spatial position and structural information data includes the adhesion relationship between the spatial positioning and structural information of each layer and the layer thickness width;

[0093] Step 1122: According to the obtained spatial position and structural information data, combined with the patient's skin laxity, fat distribution, and fat content, perform a simulation to make the facial simulated three-dimensional image match the facial anatomical structure information data to obtain a matching result.

[0094] Step 1123: Based on the matching result, obtain the defective structure anatomical image data. Divide the anatomical structure of the defective area of each defective structure anatomical image data, and partition and label the divided defective areas to accurately position the parts of each defective area anatomical data item.

[0095] In the implementation of the present invention, in Step 1121, first, different levels of information of the face are collected and organized into a data set, including:

[0096] The skin layer, the condition of the facial surface skin, such as thickness and firmness;

[0097] The subcutaneous tissue layer, the fat layer under the skin, which affects the contour and volume of the face;

[0098] The muscle layer, the distribution, thickness, and functional changes of the facial muscle tissue;

[0099] The nerve layer, which involves the distribution and function of the facial nerves and may affect facial expressions and movements;

[0100] The vascular layer, the blood vessel distribution, especially the microvascular conditions of the face.

[0101] The spatial positioning and structural information of each layer are extracted and recorded. At the same time, the mutual adhesion relationships between each layer (such as the adhesion between the skin layer and the subcutaneous tissue layer), as well as data such as layer thickness and width, are also recorded in detail. Through the accurate modeling of different levels of the face, a more comprehensive understanding of the functions of each layer and their impacts on defects can be achieved. This can help doctors or professionals understand how layers such as the skin, muscles, nerves, and blood vessels interact with each other and help diagnose the root causes of defects. The analysis of these data can better predict facial changes, especially problems such as skin laxity and fat distribution changes during the aging process.

[0102] Step 1122, based on the spatial position and structural information data obtained in Step 1121, combined with the specific facial features of the patient (such as skin laxity, fat distribution, fat content, etc.), uses three-dimensional simulation technology to simulate the face. During the simulation process, by comparing the facial anatomical structure information of the patient with the three-dimensional image, the simulation parameters are adjusted to make the simulation result conform to the actual facial structure. This process may use technologies such as deep learning and machine learning to optimize the simulation effect and ensure the accuracy of the simulation result; by combining the specific data of the patient individual, the simulation can provide accurate facial dynamics and structural performance, making the simulation process more personalized and realistic; doctors can see the facial changes in advance through the simulation result, which helps them make accurate decisions in the treatment or surgical plan and reduce risks; through the automated matching process, human errors can be reduced, ensuring that the simulation image is as consistent as possible with the actual facial anatomical structure and improving the effect of subsequent treatment.

[0103] Step 1123, based on the matching result obtained in Step 1122, forms detailed anatomical images of facial defects. These images include not only the anatomical information of each layer of the face but also show the specific manifestations of the defects (such as wrinkles, sagging, fat accumulation, muscle relaxation, etc.) at these levels; through further analysis of the defect area, the defect areas in the facial anatomical image are partitioned. For example, analyze whether a certain part of the wrinkle is related to skin laxity, fat distribution, or muscle relaxation. The anatomical structure of each area will be divided separately according to its characteristics; the divided defect areas are marked to ensure that each defect data item can accurately correspond to a specific facial anatomical part. This process is realized automatically by an algorithm, avoiding the errors of manual operation and accurately marking the specific parts of the defect area; through the accurate anatomical image and marking, doctors can clearly understand the specific position and influence range of each defect, providing more scientific data support for subsequent treatment; by partitioning and marking the defect areas, detailed anatomical analysis can be provided for each defect area, helping doctors identify its deep-seated reasons; based on the accurate positioning and anatomical data of the defects, doctors can formulate more personalized treatment or surgical plans, thereby improving the success rate and effect of treatment.

[0104] In a preferred embodiment of the present invention, in Step 1123, the anatomical structure of the defect area of each defect structure anatomical image data is divided, and the divided defect areas are partitioned and marked to accurately position the anatomical data items of each defect area, including:

[0105] Step 11231, for the defect area in each defect structure anatomical image data, through C i ={p|I(p)=1 + C(p, C j ), the adjacent pixel points are connected together to form different defect partitions;

[0106] Among them, C i represents the i-th connected region, p represents a pixel point in the image, I(p) represents the pixel vector of the pixel, and C(p, C j ) represents the connection vector between the pixel p and the region C j ;

[0107] Step 11232: Perform topological analysis on different defect partitions according to the spatial position and structural information data of the facial defect skin layer, subcutaneous tissue layer, muscle layer, nerve orientation, and blood vessel distribution, so as to divide the defect area into different structural layers and remove the defect areas with noise and connection breaks;

[0108] Step 11233: Perform hierarchical marking on different structural layers to accurately obtain the intersection points of the muscle layer, nerve orientation, and blood vessel distribution in the same defect area.

[0109] In the implementation of the present invention, in Step 11231, for the defect area in each defect structure anatomical image data, image processing technology is used to connect adjacent pixel points. This process is achieved through a "connectivity analysis" method. Usually, the continuous regions in the image are identified through algorithms, and then it is determined which pixel points belong to the same defect area; according to the connection of adjacent pixel points, the facial defect area is automatically divided into different "partitions". These partitions represent different parts of the facial defect in the image and are partitioned according to features such as spatial position, shape, color, etc.; by automatically identifying and connecting adjacent pixel points, different defect areas can be quickly and accurately divided, reducing the subjective deviation and inconsistency of manual partitioning; the automatic partitioning technology can greatly improve the analysis speed. Especially when facing complex facial images, it can quickly obtain clear defect areas, saving time for subsequent processing; through the connectivity analysis of adjacent pixel points, isolated noise points in the image can be effectively avoided, ensuring that the divided defect areas are more representative and accurate.

[0110] Step 11232 Topological analysis is a mathematical method used to analyze the structural relationships in an image. Especially when dealing with image segmentation, it can help to more accurately define the relative positions and connectivity between regions. In this step, the system utilizes the spatial position and structural information data of facial defects (including information such as the skin layer, subcutaneous tissue layer, muscle layer, etc.) for topological analysis. In facial anatomical images, there may be some pseudo-defect regions or image noises, especially errors caused by lighting changes or texture differences. Through topological analysis, the system can identify and remove these untrue noise regions. At the same time, for defect regions with broken connections, topological analysis can reconnect the broken parts to ensure the integrity of each defect region. Through topological analysis, it is also possible to accurately divide the defect regions into different anatomical levels (such as the skin layer, fat layer, muscle layer, nerve layer, and blood vessel layer) based on the spatial positioning and structural information between different tissue layers, which helps to conduct a detailed analysis of the depth and level of facial defects. Through topological analysis, it can effectively eliminate irrelevant noises and pseudo-defects in the image, ensuring the purity and accuracy of the segmentation results. For defect regions with breaks or irregular shapes, topological analysis can repair these missing regions, making the defect regions more coherent and consistent. By distinguishing different anatomical layers, it is possible to better understand the underlying causes of facial defects. For example, some defects may be caused by skin laxity, while others may involve issues such as fat distribution or muscle relaxation.

[0111] Step 11233 Based on Steps 11231 and 11232, the segmented facial defect regions are marked layer by layer. The purpose of this step is to clearly mark the specific tissue layer (such as the skin layer, muscle layer, nerve layer, blood vessel layer) to which each defect region belongs. The different tissue layers of the face are usually intertwined, especially there are many intersection points between layers such as muscles, nerves, and blood vessels. By analyzing the facial anatomical data and spatial positions, these intersection points can be accurately located, that is, the intersection points of different layers within the same defect region are marked. Through this accurate marking, it can be clearly known which regions involve multi-layer structural intersections. The layer-by-layer marking of different structural layers can clearly display the layer information of each defect region, enabling subsequent treatments to more precisely intervene in specific tissue layers. By accurately marking the intersection points, it can help doctors or professionals understand the interrelationships between different layer structures and their roles in the defect regions. For example, the intersection points of the nerve and blood vessel layers may be related to the sensory and motor functions of the face, so special attention needs to be paid during treatment. The accurate positioning of the intersection points is crucial for formulating personalized treatment plans. Especially during facial plastic surgery, anti-aging treatments, etc., accurate layer analysis can help doctors better control the depth and scope of treatment and avoid damaging important structures (such as nerves and blood vessels).

[0112] In a preferred embodiment of the present invention, step 11233 further includes:

[0113] Step 112331: Detect the shape of the acquired facial defect to obtain two defect feature shapes, namely concave and convex, and respectively obtain the spatial volume of the facial defect through and to obtain the spatial volume of the facial defect; where V1 represents the concave feature spatial volume, V2 represents the convex feature spatial volume, h1 represents the depth of the concave, h2 represents the height of the convex defect feature, r1 represents the inner diameter of the concave; r2 represents the bottom inner diameter of the convex; L represents the top width of the convex.

[0114] Step 112332: According to the spatial volume of the facial defect, combined with the patient's skin laxity, fat distribution and fat content, through G=(V×ρ 材料 )×(1 + k), to obtain the content of the injectable micro - plastic material required for the facial defect; where G represents the value of the required injectable micro - plastic material content, V represents the concave feature spatial volume V1 or the convex feature spatial volume V2, ρ 材料 represents the density of the injection material, and k represents the facial absorption rate correction factor.

[0115] Step 112333: According to the intersection points of the muscle layer, nerve direction and blood vessel distribution in the same defect area, accurately inject the micro - plastic material into the patient's facial defect area, and through the display of an animated simulation diagram, enable the patient to intuitively receive the surgical process and interact with the doctor for feedback, so as to avoid surgical risks.

[0116] In the implementation of the present invention, step 112331 accurately measures the face through advanced image - processing technology or 3D scanning technology. By detecting the concave and convex features of the facial defect, the system can quantitatively analyze the irregular areas of the face and further extract the spatial volume of these defect areas. Specifically, the shape and volume of the concave and convex can reflect the depth and severity of the patient's facial defect; through high - precision measurement of the defect area, the doctor can accurately determine the specific location that needs to be repaired, avoiding errors; it can quantify the facial defect, avoid relying solely on subjective observation, improve the accuracy of the surgery; and customize the treatment plan for each patient, and select the appropriate micro - plastic material according to different defect volumes.

[0117] Step 112332 will further calculate the specific content of the aesthetic injection material to be injected by combining factors such as the spatial volume of facial defects, the patient's skin laxity, fat distribution, and fat content. Skin laxity determines the filling requirements of the material, while fat distribution and fat content affect the uniform distribution and effect of the injection volume; by considering the characteristics of the skin and fat layers, doctors can accurately determine the injection volume of the aesthetic injection material according to individual differences to ensure a natural effect; over-injection may cause unnatural appearance or complications, and accurate calculation of the injection volume can avoid these problems; an appropriate injection volume can reduce postoperative discomfort and recovery time.

[0118] Step 112333 performs very precise injection operations based on the intersection points of the facial muscle layer, nerve layer, and blood vessel layer to avoid accidentally injuring important tissues. By using 3D animation or virtual reality technology, patients can intuitively see the surgical process, interact with the doctor, and provide real-time feedback. This simulated animation can help patients better understand the surgical process and risks, increasing patients' trust in the surgery; when injecting the aesthetic injection material, it can avoid important structures such as blood vessels and nerves, reducing the risk of postoperative complications and enhancing the safety of the surgery; through the visual simulated animation, patients can not only clearly understand the surgical process but also make adjustments based on the feedback provided by the animation; the visual process makes patients feel more at ease and reduces preoperative anxiety; the interaction between patients and doctors enables the treatment plan to be adjusted at any time to ensure the best effect.

[0119] In a preferred embodiment of the present invention, the facial aesthetic prediction model in step 14 includes:

[0120] Step 141 generates structural data from the obtained multiple facial defect feature data, multiple defect structure anatomical information data, and aesthetic treatment plan information data sets, encodes the structural data into sequence data, and trains the facial aesthetic prediction model;

[0121] Step 142 inputs the sequence data into the facial aesthetic prediction model; the facial aesthetic prediction model includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer, transmits the intermediate representation data of multiple hidden layers to the output layer, and the output layer outputs the recognition result of the aesthetic treatment plan information data corresponding to the defect structure anatomical information data representing the facial defect feature data set;

[0122] Step 143 inputs at least one newly obtained facial defect feature data into the facial aesthetic prediction model to output the prediction result of the aesthetic treatment plan information data corresponding to the defect structure anatomical information data representing the facial defect feature data.

[0123] In the implementation of the present invention, step 141 integrates facial defect feature data, defect structure anatomical information data, and microplastic surgery plan information data from multiple sources to form a dataset containing multi-dimensional information (i.e., "structural data"). These data usually include different defect features of the patient's face (such as depressions, protrusions, skin laxity, etc.), as well as the corresponding anatomical structures and treatment plans (such as injection volume, injection location, etc.). Then, this structural data is encoded into sequence data (for example, through numerical or vectorization processing) and used to train a deep learning model (such as a neural network). During the training process, the model will learn how to predict the corresponding microplastic surgery plan (such as filler type, dosage, injection location, etc.) based on the input data (facial defect features, anatomical information, etc.); by integrating multiple data sources (defect features, anatomical information, microplastic surgery plan) into a structured dataset, all relevant information can be fully utilized to improve the prediction ability of the model; through the trained neural network model, a personalized microplastic surgery plan can be automatically generated based on the input facial defect feature data, reducing manual intervention; the training of the model can give a more personalized and accurate treatment plan according to the facial features, anatomical structures, etc. of different patients.

[0124] Step 142 applies the already trained facial microplastic surgery prediction model in step 141 to actual prediction. New facial defect data (such as the patient's facial features, skin condition, etc.) is input into the model. The model includes multiple hidden layers, and each layer extracts different levels of features from the input data and finally generates a microplastic surgery plan prediction result at the output layer. Specifically, the input layer of the model receives the defect data and gradually passes it to the multi-layer hidden layers for feature extraction and processing. Each layer performs a non-linear transformation on the data to capture more complex features, and finally outputs a prediction result about the treatment plan at the output layer, such as the type, dosage, and injection location of the injected microplastic material; the multi-layer neural network can extract complex features in the data layer by layer, thus capturing the details of facial defects and making the prediction more accurate; the multiple hidden layers of the model help to process complex non-linear relationships, which can improve the understanding of the complexity of facial defects and the prediction accuracy; it can quickly make predictions on newly input facial defect data, thus giving a personalized microplastic surgery plan in real time and improving the diagnosis and treatment efficiency.

[0125] The newly obtained facial defect feature data in step 143 (such as new photos, scan data or symptom changes when the patient visits the doctor) is input into the already trained facial plastic surgery prediction model, and the model will predict the corresponding plastic surgery plan based on this new data. This means that even for new patients or new problems that occur during the treatment process, the model can make efficient predictions based on the existing training results and generate appropriate treatment plans; even if the patient's facial features change or there are different defects, the model can quickly adapt to the new input and output the corresponding plastic surgery plan prediction; by continuously inputting new data, the model can be continuously optimized, adjusted, and improve its adaptability and prediction accuracy for new facial defect data; as the patient's treatment process changes, the model can adjust the treatment plan at any time according to the new defect feature data to ensure that the treatment effect continuously reaches the best state.

[0126] In a preferred embodiment of the present invention, step 14 further includes:

[0127] Step 144, according to the prediction result of the facial plastic surgery prediction model, to obtain the corresponding plastic surgery plan information data, and compare the corresponding plastic surgery plan information data with the historical similar facial defect feature plastic surgery plan information data to obtain a comparison result;

[0128] Step 145, based on the comparison result, obtain the subtle difference data, perform a correlation analysis on the subtle difference data to obtain the subtle difference influencing factor; further iteratively update the parameters of the facial plastic surgery prediction model according to the subtle difference influencing factor to accurately obtain the prediction result of the personalized plastic surgery plan;

[0129] Step 146, the subtle difference influencing factor includes the patient's face shape characteristics, facial features, age, gender, and personalized needs, and adjusts the prediction result of the output plastic surgery plan information data in real time to obtain a preliminary surgery plan;

[0130] Step 147, generate feedback information of the surgical effect picture and the facial change curve according to the prediction result of the plastic surgery plan information data to iteratively optimize the preliminary surgery plan and obtain the final plastic surgery plan.

[0131] In the implementation of the present invention, in step 144, the facial features of the patient are predicted according to a pre-trained facial microplastic surgery prediction model. The model will generate a preliminary microplastic surgery plan, which is designed based on the patient's facial features (such as facial contour, proportion of facial features, skin quality, etc.); then, this preliminary microplastic surgery plan is compared with the microplastic surgery plans of historical patients with similar facial defect features. Historical data usually comes from past patient cases and can provide reference for the current individual; the comparison process helps to ensure the rationality of the predicted plan, and with the experience of historical data, some potential microplastic surgery plans can be discovered, thus increasing the success rate of the plan; through comparison with historical data, the plan for the current patient can also be optimized, reducing errors and inappropriate treatment plans.

[0132] After the comparison in step 145, the algorithm extracts some "subtle difference" data based on the comparison results. These data usually include some relatively detailed differences, such as slight changes in facial morphology, muscle tightness, etc., which may affect the final surgical effect; then, correlation analysis is performed on these difference data to identify "subtle difference influencing factors" that have a significant impact on the surgical effect. These factors can be minor differences in facial structure or age and gender factors. Based on these influencing factors, the parameters of the facial microplastic surgery model are further updated and iterated to make the prediction results more accurate and personalized.

[0133] This step can ensure that the personalized surgical plan more precisely adapts to the patient's unique facial features, thus avoiding blindly applying historical data; by analyzing the subtle differences, some factors that may affect the effect can be identified, further improving the surgical success rate and reducing possible side effects or unsatisfactory results.

[0134] In step 146, based on step 145, the personalized needs of the patient are further integrated, including factors such as facial contour, proportion of facial features, age, and gender, combined with the previously extracted subtle difference influencing factors, to adjust the prediction result of the microplastic surgery plan. These adjustments are not limited to the adjustment of facial features, but also include the patient's personalized aesthetic needs, such as whether a younger, more natural or more three-dimensional effect is required; through real-time adjustment and optimization, the personal needs of the patient can be better met, enhancing the personalization and customization of the plan, and improving patient satisfaction; carefully considering factors such as age and gender makes the surgical plan more in line with physiological characteristics, avoiding overly extreme or unnatural facial aesthetics treatments.

[0135] In step 147, based on the prediction results of the microplastic surgery plan, the system will generate surgical effect diagrams and facial change curves, which show the facial change effects before and after the surgery. Through the effect diagrams and change curves, patients and doctors can intuitively see the expected results of the surgical effect, which helps both parties confirm whether the surgical plan is appropriate. This process also provides a feedback mechanism to help iterate and optimize the plan, ensuring that the final plan is the most suitable, meets the patient's needs and is effective. The generation of the effect diagrams and facial change curves enables patients to intuitively feel the possible changes before the actual surgery, helping doctors and patients communicate and adjust the plan better. This feedback mechanism can continuously optimize the surgical plan, increasing the success rate of the surgery and patient satisfaction.

[0136] In a preferred embodiment of the present invention, the surgical precise operation strategy in step 16 includes:

[0137] Step 161, customize the microplastic surgery path according to the differences in the intersection points of the muscle layer, nerve direction and blood vessel distribution in the same defect area of the facial defect anatomical structure, as well as the directions of the three major facial nerves, so as to avoid the areas where nerves are densely intertwined.

[0138] In the implementation of the present invention, before the surgery in step 161, the doctor will conduct a detailed analysis based on the anatomical structure of the patient's facial defect area. Specifically, analyze the intersection points of the muscle layer, nerve layer and blood vessel layer in the analysis area, as well as the directions of the three major facial nerves (such as the trigeminal nerve, facial nerve, etc.). The facial muscles, nerves and blood vessels are densely distributed, and great care needs to be taken to avoid damaging these key structures during microplastic surgery. Through 3D modeling or anatomical images, the doctor can accurately understand the anatomical structures at different levels to identify the areas where nerves are intertwined. Based on these anatomical analyses, the doctor will customize a microplastic surgery path to ensure avoiding the areas where nerves are densely intertwined, thereby minimizing the risk of nerve injury during the surgery. By accurately analyzing and customizing the surgery path, the risk of damage to nerves, blood vessels and muscles can be significantly reduced, avoiding postoperative complications and sequelae such as nerve paralysis or muscle function impairment. Improve the safety of the surgery, especially in sensitive areas such as the face, ensure that the microplastic surgery is more delicate and effective, and reduce the possibility of postoperative discomfort or failure.

[0139] As Figure 2 shown, an intelligent simulation system for facial microplastic surgery operations includes:

[0140] Data acquisition module: It is used to acquire patient facial data, various facial defect feature data, and various defect structure anatomical information data;

[0141] Preprocessing module: It is used to perform normalization, feature extraction and encoding processing on each data;

[0142] Simulation module: It is used to fuse the processed data and perform virtual simulation to obtain various micro - plastic surgery plan information data;

[0143] Model construction module: It is used to construct a facial micro - plastic prediction model from facial data, various facial defect feature data, various defect structure anatomical information data, and various micro - plastic surgery plan information data;

[0144] Scheme prediction module: It is used to enable the facial micro - plastic prediction model to output a prediction result of the micro - plastic surgery plan information data corresponding to the defect structure anatomical information data of the facial defect feature data;

[0145] Feedback and evaluation module: It is used to obtain a personalized facial micro - plastic prediction plan based on the prediction result of the facial micro - plastic prediction model; and obtain a precise surgical operation strategy based on the facial micro - plastic prediction plan for doctor - patient interaction feedback and evaluation of the micro - plastic postoperative effect.

[0146] When the functions of the above - mentioned modules are implemented in the form of software function units and used as independent products, they can be stored in a computer - readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this 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 enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical discs, etc., which can store program codes.

[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 well - known general - purpose intelligent device. Therefore, the object of the present invention can also be achieved only by providing a program product containing program codes for implementing the method or system. That is to say, 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 well - known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, obviously, each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent solutions of the present invention. And the steps of performing the above - mentioned series of processes can naturally be executed in chronological order according to the described order, but it is not necessary to execute them in chronological order. Some steps can be executed in parallel or independently of each other.

Claims

1. An intelligent simulation method for facial micro plastic surgery, characterized in that: include: Acquire patient facial data, various facial defect feature data, and various defect structure anatomical information data; And normalize, extract features and encode each data; The processed data are integrated and virtually simulated to obtain a variety of micro-plastic surgery plan information data; facial data, a variety of facial defect feature data, a variety of defect structure anatomical information data and a variety of micro-plastic surgery plan information data are constructed into a facial micro-plastic surgery prediction model; The facial micro-surgery prediction model can output a prediction result of the micro-surgery solution information data corresponding to the defect structure anatomical information data of the facial defect feature data; Based on the prediction results of the facial micro-surgery prediction model, a personalized facial micro-surgery prediction solution is obtained; Based on the facial micro-surgery prediction plan, precise surgical operation strategies are obtained to provide interactive feedback between doctors and patients and evaluate the postoperative effects of micro-surgery.

2. The intelligent simulation method for facial micro plastic surgery according to claim 1, characterized in that: Acquire patient facial data and construct a three-dimensional facial model of the patient; analyze the three-dimensional facial model to obtain a variety of facial defect feature data and compile them into a facial defect feature data set; The plurality of facial defect feature data are positioned and marked to obtain a defect marking data set; and each facial defect feature data item in the defect marking data set is structurally dissected to obtain a defect structural anatomical information data set.

3. The intelligent simulation method for facial micro plastic surgery according to claim 2, characterized in that: The three-dimensional facial model is analyzed to obtain a variety of facial defect feature data to form a facial defect feature data set, including: Acquire three-dimensional facial information data of the patient based on the patient's facial data; pre-process the three-dimensional facial information data to acquire pre-processed three-dimensional facial information; Performing point cloud or mesh processing on the preprocessed three-dimensional facial information to construct a three-dimensional facial model; extracting facial features from the three-dimensional facial model to obtain facial feature data; Performing facial defect detection on the facial feature data to obtain facial defect feature data; and aggregating different types of facial defect feature data to obtain a facial defect feature data set.

4. The intelligent simulation method for facial micro plastic surgery according to claim 3, characterized in that: Performing structural dissection on each facial defect feature data item in the defect labeling data set to obtain a defect structural anatomical information data set, including: The defect structure anatomical information data set includes the spatial position and structural information data of the facial defect skin layer, subcutaneous tissue layer, muscle layer, nerve direction and blood vessel distribution; the spatial position and structural information data includes the adhesion relationship between the spatial positioning and structural information of each layer and the layer thickness width; According to the acquired spatial position and structure information data, combined with the patient's skin laxity, fat distribution and fat content, a simulation is performed to match the facial simulation three-dimensional image with the facial anatomical structure information data to obtain a matching result; Based on the matching results, the defective structure anatomical image data is obtained; the defective area anatomical structure of each defective structure anatomical image data is divided, and the divided defective area is partitioned and marked to accurately locate the location of the anatomical data items of each defective area.

5. The intelligent simulation method for facial micro plastic surgery according to claim 4, characterized in that: The defective area anatomical structure of each defective structural anatomical image data is divided, and the divided defective area is partitioned and marked to accurately locate the location of the anatomical data items of each defective area, including: For each defect area in the anatomical image data of the defect structure, Connect adjacent pixels together to form different defect partitions; Among them, C i represents the i-th connected area, p represents a pixel in the image, I(p) represents the pixel vector of the pixel, C(p,C j ) represents the pixel p and the region C j connected vectors; Performing topological analysis on different defect partitions according to the spatial position and structural information data of the facial defect skin layer, subcutaneous tissue layer, muscle layer, nerve direction and blood vessel distribution, so as to divide the defect area into different structural layers and remove noise and defect areas with broken connections; Different structural layers are marked in layers to accurately obtain the intersection points of muscle layers, nerve courses and blood vessel distribution in the same defect area.

6. The intelligent simulation method for facial micro plastic surgery according to claim 5, characterized in that: The acquired facial defects are subjected to shape detection to obtain the concave and convex defect feature shapes, and the defect features are respectively detected by and To obtain the spatial volume of facial defects; where V1 represents the spatial volume of concave features, V2 represents the spatial volume of convex features, h1 represents the depth of the concave, h2 represents the height of the convex defect feature, r1 represents the inner diameter of the concave; r2 represents the inner diameter of the bottom of the convex; L represents the top width of the convex; According to the volume of facial defect space, combined with the patient's skin laxity, fat distribution and fat content, G = (V × ρ 材料 )×(1+k) to obtain the amount of injection material required for facial defects; where G represents the amount of injection material required, V represents the volume of the concave feature space V1 or the volume of the convex feature space V2, ρ 材料 represents the density of the injected material, k represents the correction factor for facial absorption rate; According to the intersecting points of the muscle layer, nerve direction and blood vessel distribution in the same defect area, the micro-surgery material is accurately injected into the defective area of ​​the patient's face. Through simulated dynamic graphics display, the patient can intuitively receive the surgical process and interact with the doctor for feedback to avoid surgical risks.

7. The intelligent simulation method for facial micro plastic surgery according to claim 6, characterized in that: The facial micro-surgery prediction model includes: Generate structural data from the acquired facial defect feature data, the acquired defect structure anatomical information data, and the micro-surgery plan information data set, and encode the structural data into sequence data to train the facial micro-surgery prediction model; The sequence data is input into the facial micro-surgery prediction model; the facial micro-surgery prediction model includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer, and the intermediate representation data of multiple hidden layers are transmitted to the output layer, and the output layer outputs the micro-surgery scheme information data recognition result corresponding to the defect structure anatomical information data representing the facial defect feature data set; At least one newly acquired facial defect feature data is input into the facial micro-surgery prediction model to output a prediction result of the micro-surgery plan information data corresponding to the defect structure anatomical information data representing the facial defect feature data.

8. The intelligent simulation method for facial micro plastic surgery according to claim 7, characterized in that: According to the prediction result of the facial micro-plastic surgery prediction model, corresponding micro-plastic surgery plan information data is obtained, and the corresponding micro-plastic surgery plan information data is compared with historical micro-plastic surgery plan information data of similar facial defect features to obtain a comparison result; Based on the comparison results, subtle difference data are obtained to perform correlation analysis on the subtle difference data to obtain subtle difference influencing factors; Further iteratively updating the facial micro-plastic surgery prediction model parameters according to the subtle difference influencing factors to accurately obtain the prediction results of the micro-plastic surgery personalized surgery plan; The subtle difference influencing factors include the patient's facial features, facial features, age, gender and personalized needs, and the output micro-surgery plan information data prediction results are adjusted in real time to obtain a preliminary surgical plan; Based on the prediction results of the micro-surgery plan information data, the surgical effect diagram and feedback information of the facial change curve are generated to iteratively optimize the preliminary surgical plan and obtain the final micro-surgery plan.

9. The intelligent simulation method for facial micro plastic surgery according to claim 8, characterized in that: The precise surgical operation strategy includes: Based on the differences in the muscle layers, nerve directions, and blood vessel distribution of the facial defect anatomy in the same defect area, as well as the directions of the three major facial nerves, the minimally invasive surgery path is customized to avoid areas with dense nerve interweaving.

10. An intelligent simulation system for facial micro plastic surgery, characterized in that: The system is provided with an electronic device including a memory, a processor, and an intelligent simulation method program for facial micro-plastic surgery operations stored in the memory and executable on the processor. When the intelligent simulation method program for facial micro-plastic surgery operations is executed by the processor, the steps of the intelligent simulation method for facial micro-plastic surgery operations as claimed in any one of claims 1 to 9 are implemented. The system includes: Data acquisition module: used to acquire patient facial data, various facial defect feature data, and various defect structure anatomical information data; Preprocessing module: It is used to normalize, extract features and encode each data; Simulation module: used to fuse the processed data and perform virtual simulation to obtain various minimally invasive surgery plan information data; Model building module: It is used to build facial micro-plastic surgery prediction model by combining facial data, multiple facial defect feature data, multiple defect structure anatomical information data and multiple micro-plastic surgery plan information data; Solution prediction module: it is used for the facial micro-surgery prediction model to output the prediction result of the micro-surgery solution information data corresponding to the defect structure anatomical information data of the facial defect feature data; Feedback evaluation module: It is used to obtain a personalized facial micro-surgery prediction plan based on the prediction results of the facial micro-surgery prediction model; obtain a precise surgical operation strategy based on the facial micro-surgery prediction plan to provide interactive feedback between doctors and patients and evaluate the postoperative effect of micro-surgery.

Images