Face anti-aging optimization management system and method

Abstract

The invention relates to the technical field of medical cosmetology, in particular to an optimal management system and method for facial anti-aging. The method comprises the following steps: acquiring a facial three-dimensional model and a dynamic expression video, extracting a strain field and identifying high and low dynamic strain areas; after the face finite element mechanical model is established, an initial filling scheme is generated through topological optimization with the maximum supporting rigidity and the limitation of the total amount of materials as constraints; implanting the scheme into the model, performing dynamic mechanical simulation, and iteratively correcting the mechanical constraint according to a simulation result until the dynamic state is natural; and finally, receiving an adjustment instruction through a human-computer interaction interface, converting the adjustment instruction into an optimization parameter, and guiding to generate a final injection plan. According to the technical scheme, collaborative optimization and fixed-point and fixed-position layered design of the skin phase, the meat phase and the bone phase are achieved, the problems that the face is loose, vertical, fat and concave are comprehensively solved, the treatment has the three-dimensional supporting effect and the dynamic natural effect, the minimally invasive biotechnology concept is met, treatment without the recovery period can be achieved, the effects of wrinkle removing, aging resisting, tightening and lifting are finally achieved, and a user is assisted to achieve the young face state.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medical cosmetology, and in particular to a face anti-aging optimization management system and method. BACKGROUND

[0002] The field of facial rejuvenation medical cosmetology aims to improve problems such as tissue volume loss, contour relaxation and wrinkle growth caused by aging through minimally invasive injection filling and other technical means. Its application has developed from simple wrinkle removal in the early stage to multi-level comprehensive treatment covering bone phase support, fat room volume reset and skin phase smoothing. However, the existing technical solutions generally rely on the personal clinical experience and subjective aesthetic judgment of the doctor, and the injection point and dose are planned manually through two-dimensional photos or static three-dimensional scanning, lacking quantitative calculation and intelligent decision support, resulting in significant uncertainty and instability of the treatment results. Specifically, the existing technology usually considers only a single factor, such as optimizing the deep support structure with a mechanical simulation algorithm while ignoring the impact of expression activity on material distribution, or avoiding high-risk areas through dynamic video analysis without building the best mechanical effect. This disconnection often results in technical contradictions such as significant contour lifting in static state but local stiffness, distortion or unevenness in dynamic expression after surgery. The root cause lies in the lack of an intelligent design system that can optimize the biomechanical performance and dynamic expression naturalness simultaneously. In addition, the huge differences in individual facial anatomical structure, tissue characteristics and muscle movement patterns further exacerbate the above problems, making it difficult for experience-dependent methods to achieve reliable and repeatable personalized precision treatment. Therefore, there is an urgent need in the field for an intelligent solution that deeply integrates biomechanical calculation, dynamic aesthetic analysis and constraint optimization algorithm, which can automatically generate an injection material spatial distribution scheme that meets both the deep support strength requirement and ensures the naturalness of the surface dynamic expression, fundamentally solving the fundamental technical problem of the existing technology that mechanical effect and dynamic performance cannot be considered simultaneously. SUMMARY

[0003] The present application provides a face anti-aging optimization management system and method, which solves the treatment uncertainty caused by relying on the doctor's experience and manual planning in the prior art, overcomes the technical deficiency of being unable to simultaneously optimize the facial static support structure and dynamic expression naturalness, realizes the transition from experience-based operation to scientific and personalized precision planning, and effectively avoids the problems of postoperative expression stiffness and contour improvement inconsistency.

[0004] To achieve the above purpose, the embodiments of the present application disclose the following technical solutions:

[0005] On the one hand, the present application discloses a face anti-aging optimization management method, comprising the following steps:

[0006] S1. Obtain a three-dimensional facial model of the target object and a series of facial dynamic video sequences under preset expressions, process the facial dynamic video sequences, extract displacement data of facial surface points under all preset expressions, calculate the facial dynamic strain field based on the displacement data, and identify the facial high dynamic strain area and the facial low dynamic strain area.

[0007] S2. Convert the facial 3D model into a facial finite element mechanical model and define its mechanical boundary conditions. With the goal of maximizing the support stiffness of the facial finite element mechanical model in the target area and the total amount of injected material as a constraint, run the topology optimization algorithm to generate an initial spatial distribution scheme of the filling material.

[0008] S3. The initial filling material spatial distribution scheme is implanted into the facial finite element mechanical model to form a composite mechanical model. The facial dynamic strain field is applied as a load to the composite mechanical model to perform dynamic mechanical simulation to simulate the facial expression dynamics after filling, and the dynamic simulation results are output.

[0009] S4. Analyze the dynamic simulation results and determine whether there are any unnatural dynamic regions. If there are, convert the unnatural dynamic regions into new mechanical constraints and feed them back to the topology optimization algorithm. Iterate through steps S2 to S4 until there are no unnatural dynamic regions in the dynamic simulation results, and generate the final spatial distribution scheme of the filling material.

[0010] S5. Visually present the final spatial distribution scheme of the filling material and its corresponding simulated dynamic effect through the human-computer interaction interface, receive the scheme adjustment instructions proposed by the user based on the simulated dynamic effect, and convert the scheme adjustment instructions into one or more calculable optimization parameters.

[0011] S6. Input the optimization parameters into the topology optimization algorithm, perform additional parameter-guided topology optimization calculations based on the final spatial distribution scheme of the filling material, and generate and output the final injection treatment plan.

[0012] On the other hand, this solution discloses an optimized management system for facial anti-aging, including:

[0013] The data acquisition module is used to acquire a 3D facial model of the target object and a series of dynamic video sequences of facial expressions under preset expressions;

[0014] The dynamic analysis module is used to process facial dynamic video sequences, extract displacement data of facial surface points under all preset expressions, calculate the facial dynamic strain field based on the displacement data, and identify the facial high dynamic strain area and the facial low dynamic strain area accordingly.

[0015] The mechanical modeling module is used to process the 3D facial model, convert it into a facial finite element mechanical model, and define the mechanical boundary conditions of the facial finite element mechanical model.

[0016] The topology optimization module is used to run a topology optimization algorithm to generate an initial spatial distribution scheme of filling material, with the optimization objective of maximizing the support stiffness of the facial finite element mechanical model in the target area and the total amount of injected material as a constraint.

[0017] The simulation verification module is used to implant the initial spatial distribution scheme of the filling material into the facial finite element mechanical model to form a composite mechanical model; and to apply the dynamic strain field of the face as a load to the composite mechanical model to perform dynamic mechanical simulation, simulate the dynamic expression after filling, and output the dynamic simulation results.

[0018] The decision iteration module is used to analyze the dynamic simulation results and determine whether there are any unnatural regions. If so, the unnatural regions are converted into new mechanical constraints and the new mechanical constraints are fed back to the topology optimization module to initiate a new round of iterative calculations until there are no unnatural regions in the dynamic simulation results. Finally, the spatial distribution scheme of the filling material is output.

[0019] The interactive processing module is used to present the final spatial distribution scheme of the filling material and the corresponding simulated dynamic effect video through a graphical human-computer interface, and to receive scheme adjustment instructions input by the user through interface controls. The scheme adjustment instructions include at least one of the following: injection point position adjustment instructions, injection dosage adjustment instructions, and injection material type adjustment instructions. The interactive processing module converts the received scheme adjustment instructions into corresponding topology optimization parameters. The topology optimization parameters include at least one of the following: spatial coordinate offset, material density constraint value, and material property constraint value.

[0020] The scheme generation module is used to input the topology optimization parameters into the topology optimization module, triggering the topology optimization module to perform additional optimization calculations based on the final spatial distribution scheme of the filling material and the topology optimization parameters, and generate and output the final injection treatment plan.

[0021] This invention relates to an optimized management system and method for facial anti-aging. By integrating facial dynamic aesthetic analysis and topology optimization algorithms, it achieves precise differentiation and synergistic treatment of the skin, flesh, and bone structures. The design utilizes a layered material distribution scheme generated by a computational mechanics model, enabling targeted intervention at different levels to comprehensively address issues such as facial sagging, thickening, and depression. The scheme employs dynamic strain field simulation to ensure natural and smooth facial expressions after treatment. Combined with minimally invasive biotechnology concepts, it not only achieves a three-dimensional, firm, and light facial contour but also provides a treatment experience with no recovery period. The final injection treatment plan is both scientific and personalized, maximizing the application efficiency of regenerative anti-aging instruments and truly realizing comprehensive anti-aging based on individual characteristics. Attached Figure Description

[0022] Figure 1 This is a flowchart of the overall method of Embodiment 1 of the present invention.

[0023] Figure 2 This is a flowchart of the method for generating an optimized scheme according to Embodiment 1 of the present invention;

[0024] Figure 3 This is a flowchart of the parameter conversion method according to Embodiment 1 of the present invention;

[0025] Figure 4 This is an overall block diagram of the system according to Embodiment 2 of the present invention;

[0026] Figure 5 This is an interaction diagram of the system in Embodiment 2 of the present invention. Detailed Implementation

[0027] Specific embodiments of the invention will now be described in detail. Although the invention is described in conjunction with these specific embodiments, it should be understood that the invention is not intended to be limited to these specific embodiments. Rather, these embodiments are intended to cover alternative, modified, or equivalent embodiments that may be included within the spirit and scope of the invention as defined by the claims. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the invention. The invention may be practiced without some or all of these specific details. In other instances, well-known processes have not been described in detail so as not to unnecessarily obscure the invention.

[0028] When used in conjunction with the terms "comprising," "method comprising," or similar language in this specification and appended claims, the singular forms "a," "some," and "the" include plural references unless the context clearly indicates otherwise. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0029] Application Overview: In existing technologies, facial rejuvenation treatments heavily rely on doctors' experience for manual planning, lacking synergistic optimization of the dynamic mechanical properties and static support structures of soft tissues. For example, this can lead to improved static contours but stiff and distorted dynamic expressions, affecting treatment effectiveness and naturalness. Without addressing these issues, treatment outcomes become unpredictable, customer satisfaction is low, and personalized precision treatment is impossible. To address these problems, this application first considers the deep integration of biomechanical calculations and dynamic aesthetic analysis. This application attempts to acquire 3D facial models and dynamic video data, extract strain fields, and construct finite element mechanical models. Combined with topology optimization algorithms, an initial material distribution scheme is generated. Then, dynamic simulation iteratively corrects the scheme's constraints to resolve the contradiction between mechanical support and dynamic naturalness. This allows for layered injection planning with synergistic optimization of the skin, flesh, and bone structures. Furthermore, a human-computer interaction mechanism further adapts to personalized needs, ultimately outputting an injection treatment plan that guides minimally invasive biotechnology implementation, achieving a three-dimensional, firm, and naturally dynamic anti-aging effect.

[0030] Example 1

[0031] An optimized management method for facial anti-aging, characterized by the following steps:

[0032] S1. Obtain a three-dimensional facial model of the target object and a series of facial dynamic video sequences under preset expressions, process the facial dynamic video sequences, extract displacement data of facial surface points under all preset expressions, calculate the facial dynamic strain field based on the displacement data, and identify the facial high dynamic strain area and the facial low dynamic strain area.

[0033] S2. Convert the facial 3D model into a facial finite element mechanical model and define its mechanical boundary conditions. With the goal of maximizing the support stiffness of the facial finite element mechanical model in the target area and the total amount of injected material as a constraint, run the topology optimization algorithm to generate an initial spatial distribution scheme of the filling material.

[0034] S3. The initial filling material spatial distribution scheme is implanted into the facial finite element mechanical model to form a composite mechanical model. The facial dynamic strain field is applied as a load to the composite mechanical model to perform dynamic mechanical simulation to simulate the facial expression dynamics after filling, and the dynamic simulation results are output.

[0035] S4. Analyze the dynamic simulation results and determine whether there are any unnatural dynamic regions. If there are, convert the unnatural dynamic regions into new mechanical constraints and feed them back to the topology optimization algorithm. Iterate through steps S2 to S4 until there are no unnatural dynamic regions in the dynamic simulation results, and generate the final spatial distribution scheme of the filling material.

[0036] S5. Visually present the final spatial distribution scheme of the filling material and its corresponding simulated dynamic effect through the human-computer interaction interface, receive the scheme adjustment instructions proposed by the user based on the simulated dynamic effect, and convert the scheme adjustment instructions into one or more calculable optimization parameters.

[0037] S6. Input the optimization parameters into the topology optimization algorithm, perform additional parameter-guided topology optimization calculations based on the final spatial distribution scheme of the filling material, and generate and output the final injection treatment plan.

[0038] In the embodiment, S1, when acquiring data, the facial 3D model needs to cover key layers such as skin, subcutaneous fat chambers, and bones. The preset expressions are selected from types that are prone to appearing unnatural, such as smiling and frowning, to ensure that the dynamic data covers key areas. When processing the video, the facial tissue displacement pattern is extracted, and the dynamic strain field is calculated to distinguish between high and low dynamic strain zones, providing a basis for topology optimization.

[0039] The subject maintains a neutral expression, and non-contact 3D scanning equipment is used to capture the anatomical structure from the forehead to the chin and from the front of the ear to the midline of the face, accurately reflecting the shape of bones, fat compartments and skin. Dynamic videos of key medical aesthetic expressions such as smiling and frowning are collected, and the videos are taken from the front and 45-degree side. Key points are located by feature point detection, facial surface point displacement data is extracted, and dynamic strain field is calculated by combining the data to distinguish high and low dynamic strain zones on the face.

[0040] S2. When constructing the facial finite element mechanical model, it is necessary to simulate the real tissue mechanical behavior based on the facial anatomical characteristics: the skin layer is simulated as an elastic membrane structure, the fat layer as a viscoelastic body, and the bones as a rigid support body, and different mechanical parameters are assigned to different structures to ensure that the model accurately reflects the facial stress deformation law. When defining the mechanical boundary conditions, the corresponding area of ​​the skull is fixed to simulate the supporting effect of the skull on soft tissue under real physiological conditions, avoiding overall displacement that does not conform to physiological logic in the simulation. Topology optimization aims to maximize the support stiffness of the target area, and the target area is clearly defined in combination with medical aesthetic needs. At the same time, the total amount of injected material is constrained by clinically safe dosage to balance the treatment effect and safety, and finally generate an initial filling material spatial distribution scheme that reflects the difference of "dense material in the deep support area and sparse material in the surface transition area".

[0041] S3. Constructing a composite mechanical model requires embedding the initial filling scheme into the finite element model, superimposing the mechanical properties of the filling material to form a natural tissue and a composite structure of the filling material. This accurately reflects the interface relationship between the two, avoiding simulation deviations such as material separation or excessive fusion with the tissue. When applying a dynamic strain field as a load, data should be loaded in stages according to the time sequence of expression onset, peak, and recovery to restore the dynamic changes of real facial expressions. The output results should include changes in facial morphology and the stress transmission path between the material and the tissue, providing a mechanical root cause reference for judging unnatural dynamic areas.

[0042] S4. Based on mechanical indicators and clinical aesthetic standards, identify areas with unnatural dynamics and set them as material-prohibited filling constraints. Restore the normal mechanical balance of these areas through an elimination method. During iteration, repeatedly verify all preset facial expression dynamic effects until no abnormalities are found, generating a final filling scheme that combines support and natural dynamics.

[0043] S5 balances professionalism and ease of use. When visually presented, it supports doctors to view the distribution of filling materials from multiple angles. The simulated dynamic effects support pausing and slowing down to observe key frame details. After receiving adjustment instructions from doctors, it transforms subjective aesthetic judgments into quantitative optimization parameters to ensure that the adjustment intentions are accurately transmitted to the algorithm.

[0044] S6. Based on the original plan, make local fine-tuning to avoid the loss of previous results due to overall reconstruction. For example, when enhancing the support of the chin, only optimize the material density of that area, while ensuring the dynamic naturalness of adjacent areas. Finally, generate an injection treatment plan that connects mechanical optimization and clinical operation. Doctors can directly perform injections according to the details of the plan.

[0045] This solution further proposes that the extraction of displacement data of facial surface points under all preset expressions in step S2 specifically includes:

[0046] S21. Extract facial feature points from each frame of the facial dynamic video sequence;

[0047] S22. Based on the facial feature points, perform spatial registration on all frame images;

[0048] S23. On the registered image sequence, the motion trajectory of all facial surface points is tracked by optical flow method to obtain the displacement data.

[0049] In an embodiment,

[0050] This step focuses on the accurate extraction of facial surface point displacement data, which is the foundation for subsequent dynamic strain field calculation and topology optimization. When extracting facial feature points for each frame in step S21, a feature point detection technology suitable for medical aesthetic scenarios must be employed. This technology needs to accurately identify points corresponding to subtle facial anatomical structures. In addition to common points such as the corners of the eyes, mouth, nose, and brow peaks, it should also include points closely related to the filling effect, such as the starting point of crow's feet, the apex of nasolabial folds, and the highest point of the cheekbones. These points can more comprehensively reflect the facial movement state. During the extraction process, image preprocessing is required, such as removing light interference and smoothing image noise, to ensure that the feature point positioning error is controlled within a clinically acceptable range, avoiding the impact of point deviation on the accuracy of subsequent displacement data.

[0051] Spatial registration in step S22 is crucial for ensuring the comparability of displacement data. During the operation, a neutral expression frame must be selected as the reference frame. A neutral expression indicates no significant facial muscle contraction, serving as a stable reference. Then, facial feature points from other expression frames are aligned with their corresponding points in the reference frame. Coordinate transformation is used to eliminate positional discrepancies between different frames, ensuring that the facial coordinates of all frames are in the same spatial coordinate system. The alignment accuracy of key points, such as the relative positional relationship between the corners of the eyes and the brow peaks, must be repeatedly verified during registration. If deviations occur, the registration parameters must be readjusted to ensure that the registered image sequence accurately reflects the relative movement of facial surface points.

[0052] When using optical flow to track the motion trajectory of facial surface points in step S23, motion vectors need to be calculated based on pixel grayscale changes between adjacent frames. This method can capture subtle movements in non-feature point regions, such as minute stretches of facial skin. During tracking, reasonable optical flow calculation parameters need to be set to balance computational accuracy and efficiency. Simultaneously, the tracking results need to be continuously verified. If the motion trajectory of a surface point suddenly jumps, it needs to be determined as a tracking anomaly, and the displacement data of that point needs to be corrected using interpolation. The final displacement data must include the three-dimensional coordinate changes of each facial surface point under all preset expressions, providing complete motion information for subsequent dynamic strain field calculations.

[0053] This solution further proposes that the calculation of the facial dynamic strain field in step S2 is as follows: inputting the displacement data into a pre-trained strain field calculation model, and the pre-trained strain field calculation model outputs the facial dynamic strain field.

[0054] In this embodiment, this step calculates the dynamic strain field of the face using a pre-trained model. The core principle is to ensure the accuracy and efficiency of strain calculation through a data-driven approach. The construction of the pre-trained model requires a large amount of clinical and biomechanical experimental data. First, facial displacement data of individuals of different ages, genders, and facial shapes are collected. Simultaneously, real strain field data under corresponding facial expressions is acquired using optical measurement equipment, forming a dataset for model training. During training, the displacement data of facial surface points is used as input, and the real strain field data is used as output. A deep learning model is employed for training. The model parameters need to be continuously adjusted during training until the error between the model's output strain field and the real strain field meets the requirements of clinical aesthetic medicine.

[0055] In practical applications, when inputting extracted facial surface point displacement data into a pre-trained model, the data must first be standardized to ensure its format matches the input format used during model training. This includes standardizing coordinate units and adjusting data dimensions. The facial dynamic strain field output by the model must include multi-dimensional information such as normal strain and shear strain. Normal strain reflects the degree of stretching or compression of the tissue, while shear strain reflects the shear deformation of the tissue. This strain information must correspond to the facial anatomy. For example, a larger normal strain in the perioral region indicates significant stretching during facial expressions, requiring the selection of more elastic materials for subsequent filling. A smaller shear strain in the forehead region indicates more stable movement, allowing for an increase in material stiffness to enhance support. This precise strain field analysis provides a scientific basis for optimizing subsequent filling schemes.

[0056] This solution further proposes that the criterion for judging whether there is a dynamic unnatural region in step S7 is: in the dynamic simulation of the composite mechanical model, the equivalent strain or equivalent stress of a local area exceeds the preset physiological threshold of the facial tissue type to which it belongs.

[0057] In this embodiment, the core of identifying unnatural dynamic areas is to establish a criterion that conforms to the physiological characteristics of facial tissues. This criterion needs to combine the requirements for natural facial expressions in medical aesthetics clinics with biomechanical principles. First, it is necessary to clarify the facial tissue types corresponding to different areas in the composite mechanical model. For example, the periorbital area is mainly composed of skin and subcutaneous fat, the periorbital area contains more muscle tissue, and the forehead area contains the frontalis muscle and subcutaneous fat. Different tissue types have different biomechanical tolerances. Therefore, it is necessary to set specific equivalent strain and equivalent stress preset physiological thresholds for each tissue. These thresholds need to be obtained through a large number of biomechanical experiments, such as stretching and compression experiments on isolated skin and muscle tissue to determine their maximum strain and stress values ​​without functional or morphological abnormalities. At the same time, combined with the strain range corresponding to natural facial expressions after filling in clinical cases, a reasonable threshold range is finally determined.

[0058] During dynamic mechanical simulation, it is necessary to monitor the equivalent strain and equivalent stress of each unit in the composite mechanical model in real time, and compare the monitored values ​​with the preset physiological thresholds of the corresponding tissues. If the equivalent strain or equivalent stress of a certain area exceeds the threshold, it indicates that the force or deformation of that area during facial expression movements exceeds the normal physiological range. After filling, it is prone to unnatural phenomena such as stiffness and distortion. For example, if the equivalent strain of a certain area around the mouth exceeds the physiological threshold of the skin tissue during simulated smiling, the actual filling of this area may result in restricted mouth movement, manifesting as an unnatural smile. During the judgment process, it is necessary to verify each expression type one by one to ensure coverage of all preset expressions and avoid missing any unnatural areas under a certain expression. At the same time, it is necessary to record the specific location of the unnatural area, the strain exceeding the threshold, and the stress value, so as to provide detailed basis for subsequent adjustment of constraint conditions.

[0059] This scheme further proposes that, in step S7, the dynamic unnatural region is converted into a new mechanical constraint condition. Specifically, in the design variable space of the topology optimization algorithm, the spatial coordinate points corresponding to the dynamic unnatural region are marked as material-prohibited filling regions.

[0060] In this embodiment, the dynamic unnatural region is transformed into a new mechanical constraint condition. Essentially, this involves limiting the material filling range to prevent the filling material from interfering with the normal movement of the unnatural region. This process needs to be closely integrated with the design logic of the topology optimization algorithm. The design variable space of the topology optimization algorithm is based on the elements of the facial finite element mechanical model. The design variable of each element represents whether the element is filled with material. Therefore, the transformation operation first needs to determine the finite element elements corresponding to the dynamic unnatural region. By matching the spatial coordinates of the unnatural region with the element coordinates of the finite element model, all elements containing the unnatural region are accurately located. These elements are the objects that need to be restricted from filling in the future.

[0061] Next, constraints were set in the topology optimization algorithm to explicitly mark these units as areas where material filling is prohibited. Specifically, the design variable values ​​for these units were limited to 0, meaning the algorithm could not allocate filling material to these units during iterative calculations. The principle behind this operation is that the stiffness of filling materials is usually higher than that of natural facial tissue. Filling unnatural areas with material increases the overall stiffness of that area, hindering deformation during facial expressions. By prohibiting filling, the area retains the original mechanical properties of the tissue and can participate normally in facial expressions. For example, if a certain area around the eyes is identified as a dynamically unnatural area, prohibiting filling material in this area allows the skin around the eyes to maintain normal stretching and contraction during blinking, avoiding stiffness in eyelid movement. At the same time, it is necessary to ensure that the boundary transition of the prohibited filling area is natural, avoiding significant differences in mechanical properties with surrounding filled areas that could lead to new unnatural phenomena.

[0062] This scheme further proposes that the topology optimization algorithm mentioned in step S4 specifically adopts the variable density method, which uses the relative density of each finite element as the design variable.

[0063] In this embodiment, the variable density method is used for topology optimization because it allows for flexible adjustment of material distribution, catering to the differentiated needs for support and naturalness in different areas of the cosmetic procedure. The core logic of the variable density method is to use the relative density of each finite element element as a design variable. This variable ranges from 0 to 1, where 0 represents no filling material within the element, and 1 represents the element reaching its maximum density. Values ​​between 0 and 1 represent the material filling ratio. This continuous range allows for a smooth transition in material distribution, avoiding harsh boundaries between filled and unfilled areas and reducing the risk of unevenness on the face after filling.

[0064] When applying this method in medical aesthetics settings, the optimization direction of design variables needs to be set in conjunction with the mechanical requirements of the target area. For example, for the problem of sagging cheeks, the target area is the finite element unit corresponding to the cheekbone. Optimization requires increasing the relative density of these units to enhance support stiffness and improve the sagging shape. For filling fine lines around the eyes, the target area is the surface unit around the eyes. The relative density of the control unit needs to be kept at a low level to ensure that the material has good flexibility and does not affect eye expression movements. At the same time, the variable density method needs to work in conjunction with the total amount of injected material constraint. By setting an upper limit for the sum of the products of the relative density of all units and the unit volume, the total amount of material used can be controlled. This upper limit needs to refer to the clinically safe dosage, such as determining it based on the skin volume and blood supply of different areas of the face, to avoid increasing the risk of complications due to excessive material. During the algorithm operation, the relative density of each unit needs to be continuously adjusted through iterative calculations until the support stiffness of the target area reaches its maximum value under the premise of satisfying the total material constraint. The final generated initial filling material spatial distribution scheme can reflect the characteristics of "allocation on demand", laying the foundation for subsequent optimization.

[0065] This solution further proposes that the method includes the following steps before step S1: acquiring the three-dimensional facial model and acquiring the facial dynamic video sequence.

[0066] In this embodiment, acquiring a facial 3D model and dynamic video sequence before step S1 is a prerequisite for ensuring the personalization and precision of subsequent optimization solutions. This acquisition process must strictly adhere to clinical operating procedures in aesthetic medicine, balancing data accuracy and customer experience. A non-contact optical 3D scanner is used to acquire the facial 3D model. This device does not require contact with the face, avoiding pressure or irritation to the skin, while offering high scanning accuracy. Before the procedure, the target subject must remove makeup and facial accessories, ensuring that hair does not obscure the facial contours. During the scan, the subject must maintain a neutral expression, with the head fixed on a dedicated support to prevent head movement that could lead to data deviation. The scanning area must cover the entire area from the hairline to the bottom of the chin and from the left ear to the right ear, ensuring the model includes all anatomical structures related to filling, such as the skull, fat compartments, and skin. After scanning, the model must be preprocessed to remove noise points and smooth the surface, ensuring the model accurately reflects the true facial morphology.

[0067] When capturing dynamic facial video sequences, a high-definition camera should be used to clearly capture subtle facial movements, such as the minute contractions of the orbicularis oculi muscle. The shooting environment must ensure uniform lighting, avoiding direct sunlight or shadows; a ring light can be used to provide a soft light source. Before shooting, the subject should be instructed to familiarize themselves with the standard movements of the preset facial expressions. During shooting, a dual-camera setup should be used simultaneously, with the frontal camera capturing movement along the midline of the face and the side camera capturing dynamic changes in the facial contours, ensuring complete data encompassing the onset, movement, and recovery phases of the expression. After shooting, the video sequence must be filtered to remove invalid frames caused by blinking or head movements, ensuring the accuracy of subsequent data processing.

[0068] This plan further proposes that the injection treatment planning scheme includes the three-dimensional coordinates of the injection point, the injection depth, the type of injection material, and the injection dosage.

[0069] The injection treatment plan must fully meet the needs of clinical cosmetic procedures, with all aspects precise and clearly defined to directly guide the doctor in the injection process. The three-dimensional coordinates of the injection points must be determined based on the final spatial distribution plan of the filler material. When marking specific points on the facial 3D model, it is necessary to consider the facial anatomy and the filling target—for example, the injection point for improving nasolabial folds should be selected in the fat compartment area on the outer side of the nasolabial folds, with coordinates accurate to the millimeter level. At the same time, the corresponding anatomical layer of the point should be marked to avoid injecting too deeply, damaging nerves and blood vessels, or injecting too superficially, causing unevenness on the skin surface.

[0070] The injection depth needs to be determined comprehensively based on the injection site and the type of filler material. Different facial areas have significantly different anatomical depths. For example, the skin around the eyes is thinner, so the injection depth is usually controlled at the junction of the dermis and subcutaneous fat layer to avoid penetrating the skin or damaging the nerves around the eyes. The chin area requires deep support, and the injection depth needs to reach the periosteum to ensure the material provides stable support. The choice of injectable material type must match the dynamic and supportive needs of the area. Highly dynamic areas require materials with low elastic modulus and good flexibility, which can deform synchronously with muscle movement to avoid facial stiffness. Deep support areas require materials with high elastic modulus and strong support, which can maintain the support effect for a long time and improve facial contour.

[0071] The injection dosage needs to be determined based on the material requirements at each injection point in the spatial distribution plan of the filler material, and adjusted according to the individual's tissue volume loss. For example, the dosage can be appropriately increased for individuals with severe facial fat loss; for individuals with good skin elasticity, the dosage needs to be controlled to avoid overfilling. The dosage must be marked to the nearest 0.1 ml, and the injection sequence and injection speed at each injection point must be noted to ensure that the injection treatment plan is both scientific and practical, effectively improving the efficacy and safety of cosmetic treatments.

[0072] Example 2

[0073] A facial anti-aging optimization management system, the system comprising:

[0074] The data acquisition module is used to acquire a 3D facial model of the target object and a series of dynamic video sequences of facial expressions under preset expressions;

[0075] The dynamic analysis module is used to process facial dynamic video sequences, extract displacement data of facial surface points under all preset expressions, calculate the facial dynamic strain field based on the displacement data, and identify the facial high dynamic strain area and the facial low dynamic strain area accordingly.

[0076] The mechanical modeling module is used to process the 3D facial model, convert it into a facial finite element mechanical model, and define the mechanical boundary conditions of the facial finite element mechanical model.

[0077] The topology optimization module is used to run a topology optimization algorithm to generate an initial spatial distribution scheme of filling material, with the optimization objective of maximizing the support stiffness of the facial finite element mechanical model in the target area and the total amount of injected material as a constraint.

[0078] The simulation verification module is used to implant the initial spatial distribution scheme of the filling material into the facial finite element mechanical model to form a composite mechanical model; and to apply the dynamic strain field of the face as a load to the composite mechanical model to perform dynamic mechanical simulation, simulate the dynamic expression after filling, and output the dynamic simulation results.

[0079] The decision iteration module is used to analyze the dynamic simulation results and determine whether there are any unnatural regions. If so, the unnatural regions are converted into new mechanical constraints and the new mechanical constraints are fed back to the topology optimization module to initiate a new round of iterative calculations until there are no unnatural regions in the dynamic simulation results. Finally, the spatial distribution scheme of the filling material is output.

[0080] The interactive processing module is used to present the final spatial distribution scheme of the filling material and the corresponding simulated dynamic effect video through a graphical human-computer interface, and to receive scheme adjustment instructions input by the user through interface controls. The scheme adjustment instructions include at least one of the following: injection point position adjustment instructions, injection dosage adjustment instructions, and injection material type adjustment instructions. The interactive processing module converts the received scheme adjustment instructions into corresponding topology optimization parameters. The topology optimization parameters include at least one of the following: spatial coordinate offset, material density constraint value, and material property constraint value.

[0081] The scheme generation module is used to input the topology optimization parameters into the topology optimization module, triggering the topology optimization module to perform additional optimization calculations based on the final spatial distribution scheme of the filling material and the topology optimization parameters, and generate and output the final injection treatment plan.

[0082] This system achieves intelligent operation across the entire process from data acquisition to solution generation through the collaborative work of its various modules. The functional design of each module must closely align with the clinical needs of aesthetic medicine. The data acquisition module, as the foundation, must ensure high accuracy in acquiring 3D facial models and dynamic video sequences to provide a reliable data source for subsequent analysis and modeling. After acquisition, the data must be automatically transmitted to the dynamic analysis and mechanical modeling modules to avoid data errors caused by manual transmission.

[0083] The core task of the dynamic analysis module is to extract dynamic information. When processing dynamic video sequences, it first locates key anatomical points through feature point detection, then unifies the coordinate system through spatial registration, and finally tracks displacement data using optical flow. The entire process needs to be automated, while also providing data verification functions. If the feature point extraction error of a certain frame is too large, the module needs to automatically mark it and reprocess it. The calculated facial dynamic strain field needs to be displayed in a visual form, such as using different colors to mark high and low strain areas, so that doctors can intuitively understand the dynamic characteristics of the face. Then, the strain field data and strain area information are transmitted to the simulation verification module and the topology optimization module, respectively.

[0084] When constructing a facial finite element mechanical model, the mechanical modeling module automatically generates a mesh based on the facial anatomy, automatically increasing the mesh density in high-dynamic strain zones. Simultaneously, it automatically defines mechanical boundary conditions based on the anatomical location of the skull, eliminating the need for manual intervention and ensuring modeling efficiency and accuracy. Once the model is built, it is transferred to the topology optimization and simulation verification modules, providing a mechanical basis for subsequent calculations.

[0085] The topology optimization module needs to combine the strain zone information provided by the dynamic analysis module and the finite element model provided by the mechanical modeling module to run the algorithm with the goal of maximizing the support stiffness of the target area and controlling the total amount of material. After generating an initial scheme, it is transmitted to the simulation verification module. After receiving new constraints from the decision iteration module, it needs to automatically adjust the algorithm parameters and recalculate. After receiving optimization parameters from the interactive processing module, it needs to perform local optimization based on the original scheme to ensure that the scheme can continuously adapt to dynamic needs and the doctor's adjustment intentions.

[0086] After receiving the initial scheme, the simulation verification module automatically merges it with the finite element model to construct a composite mechanical model, and then loads a dynamic strain field to perform dynamic mechanical simulation. The simulation process needs to simulate the motion process of all preset expressions, and the output results need to include dynamic video and strain and stress data reports, which are transmitted to the decision iteration module for analysis and judgment.

[0087] When the decision iteration module analyzes the simulation results, it needs to automatically compare the strain, stress values ​​and physiological thresholds of each region to determine whether there are any dynamically unnatural regions. If so, it will automatically convert them into material non-filling constraints and transmit them to the topology optimization module to trigger iteration until there are no abnormal regions. Then, the final filling scheme will be transmitted to the interactive processing module.

[0088] The interactive processing module needs to provide a user-friendly human-computer interaction interface. When displaying the 3D visualization of the filling scheme, it should support doctors to rotate and zoom the model to view details. When playing simulated dynamic effects, it should support viewing facial expression changes frame by frame. After receiving the doctor's adjustment instructions, it should automatically parse the instruction type and content, convert them into topology optimization parameters, and transmit them to the scheme generation module.

[0089] After receiving the parameters, the treatment plan generation module automatically triggers the topology optimization module to perform additional optimizations, ultimately generating a treatment plan that includes injection points, depths, materials, and dosages. This plan can be exported as a clinical operation document for direct guidance of injection treatments. Through seamless collaboration between modules, the entire system achieves intelligent and personalized design of cosmetic filler treatment plans.

[0090] This solution further proposes that the data acquisition module includes a 3D scanner and a high-definition camera. The 3D scanner is used to acquire a 3D facial model, and the high-definition camera is used to acquire a dynamic video sequence of the face. The interaction processing module includes an instruction parsing unit and a parameter mapping unit. The instruction parsing unit is used to parse the scheme adjustment instructions input by the user and extract the adjustment elements. The parameter mapping unit is used to convert the adjustment elements into numerical optimization parameters that the topology optimization module can recognize.

[0091] The selection of equipment and the design of operation for the data acquisition module must balance accuracy and adaptability to medical aesthetic scenarios. A non-contact optical scanner is selected for the 3D scanner. This device uses the structured light scanning principle to quickly acquire facial 3D data. Equipment calibration is required before scanning. During the scanning process, the scanning progress and data integrity must be displayed in real time. After scanning, noise points and redundant data on the model surface are automatically removed to generate a smooth 3D model.

[0092] High-definition cameras need to have high resolution and high frame rate characteristics to ensure the capture of clear dynamic images; during shooting, a dedicated bracket should be used to fix the camera position to avoid camera shaking during the shooting process, and a ring light should be used to provide a uniform light source to eliminate facial shadows and ensure that facial features are clearly distinguishable in the image; after shooting, the camera will automatically perform frame synchronization processing on the video sequence to ensure that the timeline of the front and side videos is consistent, which will facilitate subsequent multi-angle analysis.

[0093] The two units of the interactive processing module need to achieve accurate conversion between instructions and parameters. When the instruction parsing unit receives adjustment instructions input by the doctor through interface controls, it needs to be able to identify different types of instructions.

[0094] When the parameter mapping unit converts the parsed adjustment elements into topology optimization parameters, a fixed mapping relationship needs to be established, with the position adjustment corresponding to the spatial coordinate offset. The converted parameters must conform to the input format of the topology optimization algorithm to ensure that the algorithm can accurately identify and apply these parameters, achieving precise alignment between the doctor's adjustment intentions and the algorithm's optimization logic, and improving the personalization of the cosmetic surgery plan.

[0095] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them; although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications can still be made to the specific implementation methods of the present invention or equivalent substitutions can be made to some technical features without departing from the spirit of the technical solutions of the present invention, and all such modifications should be covered within the scope of the technical solutions claimed in the present invention.

Claims

1. Method for the optimized management of facial anti-aging, characterized in that, The method comprises the following steps: S1, obtaining a three-dimensional face model of a target object and a series of face dynamic video sequences under a series of preset expressions, processing the face dynamic video sequences, extracting displacement data of face surface points under all preset expressions, calculating a face dynamic strain field based on the displacement data, and identifying a face high dynamic strain region and a face low dynamic strain region; S2, converting the three-dimensional face model into a face finite element mechanics model and defining a mechanics boundary condition thereof, taking maximizing support stiffness of the face finite element mechanics model in a target region as a target and taking total amount of injected material as a constraint, and running a topology optimization algorithm to generate an initial filling material space distribution scheme; S3, implanting the initial filling material space distribution scheme into the face finite element mechanics model to form a composite mechanics model, applying the face dynamic strain field as a load to the composite mechanics model for dynamic mechanics simulation to simulate expression dynamics after filling, and outputting a dynamic simulation result; S4, analyzing the dynamic simulation result and judging whether there is a dynamic unnatural region, if there is, converting the dynamic unnatural region into a new mechanics constraint condition and feeding back to the topology optimization algorithm, iteratively executing steps S2 to S4 until there is no dynamic unnatural region in the dynamic simulation result, and generating a final filling material space distribution scheme; S5, visualizing the final filling material space distribution scheme and a corresponding simulated dynamic effect through a human-computer interaction interface, receiving a scheme adjustment instruction based on the simulated dynamic effect, and converting the scheme adjustment instruction into one or more calculable optimization parameters; S6, inputting the optimization parameters into the topology optimization algorithm, performing additional parameter-guided topology optimization calculation on the basis of the final filling material space distribution scheme, and generating and outputting a final injection treatment planning scheme.

2. The method for the optimal management of facial anti-aging according to claim 1, characterized in that, The step S2 of extracting displacement data of face surface points under all preset expressions comprises the following steps: S21, extracting face feature points of each frame of image from the face dynamic video sequence; S22, based on the face feature points, performing spatial registration on all frames of image; S23, tracking motion trajectories of all face surface points on the registered image sequence through an optical flow method, so as to obtain the displacement data.

3. The method of claim 1, wherein, The step S2 of calculating the face dynamic strain field comprises the following step: inputting the displacement data into a pre-trained strain field calculation model, and the pre-trained strain field calculation model outputs the face dynamic strain field.

4. The method of claim 1, wherein, The judgment standard of the step S7 of judging whether there is a dynamic unnatural region is that equivalent strain or equivalent stress of a local region of the composite mechanics model in dynamic simulation exceeds a preset physiological threshold of a face tissue type to which the local region belongs.

5. The method of claim 1, wherein, The step S7 of converting the dynamic unnatural region into a new mechanics constraint condition comprises the following step: in a design variable space of the topology optimization algorithm, marking a spatial coordinate point corresponding to the dynamic unnatural region as a material prohibited filling region.

6. The method of claim 1, wherein, The step S4 of running the topology optimization algorithm specifically adopts a variable density method, and the variable density method takes relative density of each finite element unit as a design variable.

7. The method of claim 1, wherein, The method further comprises, before step S1, a step of acquiring the face three-dimensional model and acquiring the face dynamic video sequence.

8. The method for the optimal management of facial anti-aging according to any one of claims 1-7, characterized in that, The injection treatment planning scheme comprises three-dimensional coordinates of injection points, injection depths, injection material types, and injection doses.

9. An optimized management system for performing the method of any one of claims 1 to 8, characterized in that, The system comprises: a data acquisition module configured to acquire a face three-dimensional model of a target object and a face dynamic video sequence under a series of preset expressions; a dynamic analysis module configured to process the face dynamic video sequence, extract displacement data of face surface points under all preset expressions, calculate a face dynamic strain field based on the displacement data, and identify a face high dynamic strain region and a face low dynamic strain region therefrom; a mechanical modeling module configured to process the face three-dimensional model, convert it into a face finite element mechanical model, and define mechanical boundary conditions of the face finite element mechanical model; a topology optimization module configured to maximize support stiffness of the face finite element mechanical model in a target region as an optimization objective, constrain a total amount of injected material, run a topology optimization algorithm, and generate an initial filling material spatial distribution scheme; a simulation verification module configured to implant the initial filling material spatial distribution scheme into the face finite element mechanical model to form a composite mechanical model, apply the face dynamic strain field as a load to the composite mechanical model for dynamic mechanical simulation, simulate expression dynamics after filling, and output a dynamic simulation result; a decision iteration module configured to analyze the dynamic simulation result, determine whether there is a dynamic unnatural region, if there is, convert the dynamic unnatural region into new mechanical constraint conditions, and feed the new mechanical constraint conditions back to the topology optimization module to initiate a new round of iteration calculation until there is no dynamic unnatural region in the dynamic simulation result, and output a final filling material spatial distribution scheme; an interactive processing module configured to present the final filling material spatial distribution scheme and a corresponding simulated dynamic effect video through a graphical human-computer interface, receive a scheme adjustment instruction input by a user through an interface control, the scheme adjustment instruction comprising at least one of an injection point position adjustment instruction, an injection dose adjustment instruction, and an injection material type adjustment instruction; the interactive processing module converts the received scheme adjustment instruction into corresponding topology optimization parameters, the topology optimization parameters comprising at least one of a spatial coordinate offset, a material density constraint value, and a material attribute constraint value; a scheme generation module configured to input the topology optimization parameters into the topology optimization module, trigger the topology optimization module to perform additional optimization calculation based on the final filling material spatial distribution scheme and the topology optimization parameters, and generate and output a final injection treatment planning scheme.

10. The optimal management system of facial anti-aging according to claim 9, characterized in that, The data acquisition module comprises a three-dimensional scanner and a high-definition camera, the three-dimensional scanner is configured to acquire the face three-dimensional model, and the high-definition camera is configured to acquire the face dynamic video sequence; the interactive processing module comprises an instruction analysis unit and a parameter mapping unit, the instruction analysis unit is configured to analyze the scheme adjustment instruction input by the user and extract adjustment elements, and the parameter mapping unit is configured to convert the adjustment elements into numerical optimization parameters recognizable by the topology optimization module.

Images