Visual facial cosmetic surgery ai simulation method and system

Abstract

The application discloses a visual facial cosmetic surgery AI simulation method and system, relates to the technical field of image processing, and comprises an image acquisition module, a dynamic modeling module and an image reconstruction module. The image acquisition module is based on MRI water-fat separation imaging technology, acquires and identifies the three-dimensional distribution of the bone, muscle, fat and epidermis structure of the face of a patient, and constructs a muscle local elastic modulus map of the patient in a static state and in a specific expression state based on ultrasonic shear wave elasticity imaging. The dynamic modeling module is used for combining the distribution structure of the mandible region, the elastic modulus mutation and the propagation direction of the muscle region, analyzing the coupling behavior of deep fat and muscle, and indirectly inferring the fat slip path. The image reconstruction module is used for constructing a postoperative tissue slip model, correcting the area where the tissue dislocation and image folding problems occur in the image modeling in advance, generating a postoperative prediction three-dimensional image, and has the characteristics of high image authenticity.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to a visual facial cosmetic surgery AI simulation method and system. Background Technology

[0002] AI simulation in facial cosmetic surgery is a systematic approach that utilizes artificial intelligence to perform 3D modeling, structural prediction, and postoperative visual reconstruction of the face. Preoperatively, 3D images of the patient's face are acquired, and AI automatically identifies key areas such as bones, muscles, fat, and epidermis, generating a 3D postoperative prediction image based on the specific cosmetic procedure.

[0003] The mandible serves as the anchor point and lateral boundary support for facial soft tissues. Removing it is equivalent to removing a section of the bony overhang at the very bottom of the soft tissue, altering the original skin-fat-muscle tension system. When creating a 3D model of the facial skin and soft tissue after mandibular bone removal surgery, a unified AI facial reconstruction interpolation algorithm is used. However, when the muscle layers above the deep fat in the cheeks are complex, this smooth transition appears unnatural, leading to distorted postoperative predicted images. Deep fat is buried beneath the muscles, appearing as a grayscale transition zone with the masseter, buccinator, and zygomaticus muscles in images. The blurred boundaries in conventional MRI or CT images make the slippage behavior difficult to visualize. Therefore, designing a highly realistic, visualized AI simulation method and system for facial cosmetic surgery is essential. Summary of the Invention

[0004] The purpose of this invention is to provide a visual facial cosmetic surgery AI simulation method and system to solve the problems mentioned in the background art.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a visualized facial cosmetic surgery AI simulation system, comprising an image acquisition module, a dynamic modeling module, and an image reconstruction module. The image acquisition module, based on MRI water-lipid separation imaging technology, acquires and identifies the three-dimensional distribution of the patient's facial bone, muscle, fat, and epidermal structures, and constructs local muscle elastic modulus maps of the patient in static and specific facial expression states based on ultrasound shear wave elastography. The dynamic modeling module is used to combine the detected distribution structure of the mandibular region with the elastic modulus mutations and propagation directions of the muscle region to analyze the coupling behavior of deep fat and muscle and indirectly infer the fat slippage path. The image reconstruction module is used to construct a postoperative tissue slippage model, pre-correct areas where tissue misalignment and image folding problems occur in image modeling, and generate a postoperative predictive three-dimensional image.

[0006] According to the above technical solution, the image acquisition module includes an ultrasound probe, a high-speed ultrasound receiving unit, a shear wave velocity calculation module, an elastic modulus mapping module, an MRI scanning unit, and a water-lipid separation image module. The high-speed ultrasound receiving unit is electrically connected to the shear wave velocity calculation module, the shear wave velocity calculation module is electrically connected to the elastic modulus mapping module, and the MRI scanning unit is electrically connected to the water-lipid separation image module. The ultrasound probe excites transverse shear waves within the tissue using short pulses. The high-speed ultrasound receiving unit captures the signal of the shear wave propagating along the tissue. The shear wave velocity calculation module is used to calculate the propagation velocity of the wavefront at different depths and directions. The elastic modulus mapping module is used to calculate the regional elastic modulus and generate a color elastic distribution map. The MRI scanning unit is used to scan the face and construct a preoperative three-dimensional spatial model. The water-lipid separation image module performs water-lipid separation image reconstruction on the acquired multi-phase data to generate a distribution structure map of the mandibular region.

[0007] The dynamic modeling module includes an elastic mutation extraction module, a shear wave direction analysis module, a coupling modeling module, and a slip path generation module. The elastic mutation extraction module and the shear wave direction analysis module are both electrically connected to the elastic modulus mapping module. The coupling modeling module is electrically connected to the water-fat separation image module and the slip path generation module. The elastic mutation extraction module is used to analyze the spatial mutation points of the elastic modulus in the shear wave imaging image to infer the initiation source of slip. The shear wave direction analysis module is used to track the propagation path of the shear wave and infer the potential direction of fat movement. The coupling modeling module is used to construct a coupling model based on the elastic response of muscle and the static distribution of fat. The slip path generation module is used to infer the slip path of fat.

[0008] The image reconstruction module includes a complexity judgment module, a smooth transition processing module, an interpolation algorithm optimization module, and an AI facial reconstruction module. The complexity judgment module is electrically connected to the water-fat separation image module. The interpolation algorithm optimization module is electrically connected to the smooth transition processing module. The AI ​​facial reconstruction module is electrically connected to the interpolation algorithm optimization module and the slip path generation module. The complexity judgment module is used to determine the layer complexity of each region in the patient's mandibular region. The smooth transition processing module is used to pull the boundary surface to make the contour look coherent in order to deal with the problem of folds and grooves after facial image reconstruction. The interpolation algorithm optimization module is used to optimize the interpolation algorithm of the smooth transition processing after the interpolation algorithm provides the inferred fat slip path. The AI ​​facial reconstruction module is used to reconstruct the three-dimensional image of the skull after the patient's mandibular osteotomy using AI prediction methods.

[0009] A visual facial cosmetic surgery AI simulation method includes the following steps:

[0010] S1. First, visual technology and CT technology are used to obtain a scan model of the patient's facial skull and tissues. MRI water-fat separation imaging and ultrasound shear wave elastography are used to collect the fat distribution structure in the mandibular region using the MRI scanning unit. The static shear wave propagation is detected using an ultrasound probe and a high-speed receiver, and muscle elastic modulus map and fat distribution image are output.

[0011] S2. Remove the part of the mandible that needs to be removed from the facial skull and tissue scanning model, extract the shear wave propagation direction and elastic mutation point, and construct a muscle and fat coupling model by combining image structure and elastic response. Based on the changes in fat distribution images of the patient's face under specific expression states, indirectly infer the fat slippage path before and after the mandibular bone removal surgery.

[0012] S3. Based on the complexity of the muscle layer above the deep fat in the mandibular region, the predicted slip path is introduced into the tissue deformation simulation to identify and correct easily displaced and folded areas, and to optimize and smooth the contours of easily folded areas in the predicted image.

[0013] S4. Use the AI ​​image reconstruction module to generate a postoperative simulation 3D visualization image to obtain the correct fat distribution structure in the mandibular region.

[0014] According to the above technical solution, the specific method for determining the fat distribution structure in the mandibular region in S1 is as follows:

[0015] S1-1, Start the MRI scanning unit to scan the patient's face. The time-domain signal intensity of the MRI signal is: ,in For water signal strength, For fat signal intensity, The angular frequency of the water signal. The angular frequency of the fat signal. To represent the rotational change of the signal as a complex exponential function, and since the resonant frequency of fat is lower than that of water, the in-phase time point is selected. and the inverse time point This causes the water and fat signals to be in phase and out of phase at these two time points, respectively, based on the brightness of the in-phase image. Brightness of inverted image Find the sum and The size of the image is used to obtain the brightness of the water tissue image and the fat tissue image. When the value exceeds the preset value, it is determined to be adipose tissue. The water-like tissue containing muscle and soft tissue and the adipose tissue containing fat are distinguished to obtain the fat distribution structure in the mandibular region. The structure is fitted onto the facial skull and tissue scanning model and divided into multiple fat lobes according to the contour features.

[0016] According to the above technical solution, the method for obtaining the muscle elastic modulus map in step S1 is as follows:

[0017] S1-2. Using an ultrasound probe, a scan is performed above the adipose tissue area in the mandibular region, which is identified as deep fat. A high-intensity, short-duration focused ultrasound shock wave is emitted, generating a localized, minute lateral displacement within the muscle covering the adipose tissue, producing a shear wave. This shear wave propagates outward from the excitation point, passing through tissues of varying elasticity, thus increasing the shear wave velocity. Changes occur, according to the formula for the shear modulus of muscle. Organize the elastic modulus of each spatial point into a continuous field function. , representing the distribution of the elastic modulus at each point, where Using spatial coordinates, these are superimposed on the patient's facial skull and tissue scan model. According to the above technical solution, the specific method for inferring the fat slippage path in step S2 is as follows:

[0018] S2-1, After removing the portion of the mandible that needs to be shaved from the facial skull and tissue scanning model, In the detection space, coordinate points where abrupt changes occur are identified to determine the modulus gradient. Greater than the mutation threshold The point, namely the elastic abrupt change point, where , represents the vector synthesized from the partial derivative arrays in the three directions;

[0019] S2-2, Let the response time of the shear wave in the three-dimensional tissue image be... , Indicates the shear wave at point At the time of the first arrival, the propagation direction vector of the shear wave at each point is defined as: That is, along the direction where the propagation time decreases, where After normalization, the unit propagation direction is obtained. , indicating the unit direction of propagation of the shear wave at this point;

[0020] S2-3, at a certain mutation point At that point, trace the propagation direction of the shear wave. And extend to generate trajectory lines ,in For path length, The size of the mutation points in images of fat distribution in patients at rest and under specific facial expressions. Distance between the centroids of the fat lobes Proportional, that is, at a certain mutation point The more easily the fat in a given area is pulled by the covering muscle tissue, the greater the distance the fat will slide after jaw reduction surgery.

[0021] According to the above technical solution, in S3, the method for determining the complexity of muscle layers in the mandibular region is as follows: the muscles covering deep fat are divided according to the location of each fat lobe, and the muscle elasticity complexity index is used. ,in This is the point of abrupt change in the elastic modulus. This represents the total number of muscle mutation points located at the site of the fat lobe. A higher number of mutations and a greater degree of change in the direction of shear wave propagation indicate more complex muscle tissue. The muscle layers are highly complex at that time. To determine the complex threshold, it is necessary to introduce the inferred fat slip path into the interpolation algorithm.

[0022] According to the above technical solution, in step S3, optimizing and smoothing the contours of easily foldable regions in the predicted image specifically involves: based on the predicted fat slip path... The algorithm generates a fat distribution structure in the mandibular region, and simultaneously uses an interpolation algorithm to generate a fat distribution structure based on the preoperative fat distribution structure. This interpolation is performed when a certain coordinate point in the preoperative fat distribution structure is located within the adipose tissue. To the nearest slip path closest distance At that time, the interpolation weight function ,in For the set radius of influence, when hour, The weight of a point in the interpolation is determined by its distance from the sliding path. The size of a certain coordinate point The closer it is to 1, the more it tends to use the fat sliding path. The resulting fat distribution structure in the mandibular region The closer it is to 0, the more it tends to the fat distribution structure generated by the difference algorithm.

[0023] The predicted fat slip path will only be considered in areas with high muscle layer complexity in the mandibular region, and the slip path will be determined based on the coordinates of each point. The distance determines which method is used to adjust the fat distribution structure at the current coordinate point. Only when the distance to the slip path is close enough will the fat distribution be significantly adjusted. No adjustment is needed in places where fat slippage does not occur. The goal is to use interpolation algorithms in positions where image errors will not occur, and to avoid secondary deviations caused by excessive adjustment.

[0024] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: After detecting the fat distribution structure in the mandibular region using MRI water-fat separation imaging, this invention uses ultrasonic shear wave elastography combined with facial expression dynamic response to detect the elastic mutation and propagation displacement behavior of muscles. Based on this, an inverse modeling relationship of muscle and fat coupling is constructed to indirectly infer the fat slip path. In areas with complex deep fat layers in the mandible, the image is reconstructed based on the inferred fat slip path. For other locations, an interpolation algorithm is used to construct the contour, thereby correcting areas with tissue misalignment and image folding problems in image modeling. Even when the patient's mandibular region has complex layers, image distortion can still be prevented. Attached Figure Description

[0025] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0026] Figure 1 This is a schematic diagram of the overall modular structure of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Please see Figure 1 This invention provides a technical solution: a visual facial cosmetic surgery AI simulation method and system, including an image acquisition module, a dynamic modeling module, and an image reconstruction module. The image acquisition module is based on MRI water-lipid separation imaging technology to acquire and identify the three-dimensional distribution of the bone, muscle, fat, and epidermal structures of the patient's face, and to construct local elastic modulus maps of muscles under static and specific facial expression states based on ultrasound shear wave elastography. The dynamic modeling module is used to combine the detected distribution structure of the mandibular region with the elastic modulus mutation and propagation direction of the muscle region to analyze the coupling behavior of deep fat and muscle and indirectly infer the fat slippage path. The image reconstruction module is used to construct a postoperative tissue slippage model, correct the areas where tissue misalignment and image folding problems occur in advance in the image modeling, and generate a postoperative predictive three-dimensional image.

[0029] The image acquisition module includes an ultrasound probe, a high-speed ultrasound receiving unit, a shear wave velocity calculation module, an elastic modulus mapping module, an MRI scanning unit, and a water-lipid separation image module. The high-speed ultrasound receiving unit is electrically connected to the shear wave velocity calculation module, the shear wave velocity calculation module is electrically connected to the elastic modulus mapping module, and the MRI scanning unit is electrically connected to the water-lipid separation image module. The ultrasound probe excites transverse shear waves within the tissue using short pulses. The high-speed ultrasound receiving unit captures the signal of the shear wave propagating along the tissue. The shear wave velocity calculation module is used to calculate the propagation velocity of the wavefront at different depths and directions. The elastic modulus mapping module is used to calculate the regional elastic modulus and generate a color elastic distribution map. The MRI scanning unit is used to scan the face and construct a preoperative three-dimensional spatial model. The water-lipid separation image module performs water-lipid separation image reconstruction on the acquired multi-phase data to generate a distribution structure map of the mandibular region.

[0030] The dynamic modeling module includes an elastic mutation extraction module, a shear wave direction analysis module, a coupling modeling module, and a slip path generation module. The elastic mutation extraction module and the shear wave direction analysis module are both electrically connected to the elastic modulus mapping module. The coupling modeling module is electrically connected to the water-fat separation image module and the slip path generation module. The elastic mutation extraction module is used to analyze the spatial mutation points of the elastic modulus in the shear wave imaging image to infer the initiation source of slip. The shear wave direction analysis module is used to track the propagation path of the shear wave and infer the potential direction of fat movement. The coupling modeling module is used to construct a coupling model based on the elastic response of muscle and the static distribution of fat. The slip path generation module is used to infer the slip path of fat.

[0031] The image reconstruction module includes a complexity assessment module, a smooth transition processing module, an interpolation algorithm optimization module, and an AI face reconstruction module. The complexity assessment module is electrically connected to the water-fat separation image module. The interpolation algorithm optimization module is electrically connected to the smooth transition processing module. The AI ​​face reconstruction module is electrically connected to the interpolation algorithm optimization module and the slip path generation module. The complexity assessment module is used to determine the layer complexity of each region in the patient's mandibular region. The smooth transition processing module is used to pull the boundary surface to make the contour look coherent in order to deal with the problem of folds and grooves after facial image reconstruction. The interpolation algorithm optimization module is used to optimize the interpolation algorithm of the smooth transition processing after the interpolation algorithm provides the inferred fat slip path. The AI ​​face reconstruction module is used to reconstruct the three-dimensional image of the skull after the patient's mandibular osteotomy using AI prediction methods.

[0032] Includes the following steps:

[0033] S1. First, visual technology and CT technology are used to obtain a scan model of the patient's facial skull and tissues. MRI water-fat separation imaging and ultrasound shear wave elastography are used to collect the fat distribution structure in the mandibular region using the MRI scanning unit. The static shear wave propagation is detected using an ultrasound probe and a high-speed receiver, and muscle elastic modulus map and fat distribution image are output.

[0034] S2. Remove the part of the mandible that needs to be removed from the facial skull and tissue scanning model, extract the shear wave propagation direction and elastic mutation point, and construct a muscle and fat coupling model by combining image structure and elastic response. Based on the changes in fat distribution images of the patient's face under specific expression states, indirectly infer the fat slippage path before and after the mandibular bone removal surgery.

[0035] S3. Based on the complexity of the muscle layer above the deep fat in the mandibular region, the predicted slip path is introduced into the tissue deformation simulation to identify and correct easily displaced and folded areas, and to optimize and smooth the contours of easily folded areas in the predicted image.

[0036] S4. Use the AI ​​image reconstruction module to generate a postoperative simulation 3D visualization image to obtain the correct fat distribution structure in the mandibular region.

[0037] In S1, the specific method for determining the fat distribution structure in the mandibular region is as follows:

[0038] S1-1, Start the MRI scanning unit to scan the patient's face. The time-domain signal intensity of the MRI signal is: ,in For water signal strength, For fat signal intensity, The angular frequency of the water signal. The angular frequency of the fat signal. To represent the rotational change of the signal as a complex exponential function, and since the resonant frequency of fat is lower than that of water, the in-phase time point is selected. and the inverse time point This causes the water and fat signals to be in phase and out of phase at these two time points, respectively, based on the brightness of the in-phase image. Brightness of inverted image Find the sum and The size of the image is used to obtain the brightness of the water tissue image and the fat tissue image. When the value exceeds the preset value, it is determined to be adipose tissue. The water-like tissue containing muscle and soft tissue and the adipose tissue containing fat are distinguished to obtain the fat distribution structure in the mandibular region. The structure is fitted onto the facial skull and tissue scanning model and divided into multiple fat lobes according to the contour features.

[0039] In S1, the method for obtaining the muscle elastic modulus map is as follows:

[0040] S1-2. Using an ultrasound probe, a scan is performed above the adipose tissue area in the mandibular region, which is identified as deep fat. A high-intensity, short-duration focused ultrasound shock wave is emitted, generating a localized, minute lateral displacement within the muscle covering the adipose tissue, producing a shear wave. This shear wave propagates outward from the excitation point, passing through tissues of varying elasticity, thus increasing the shear wave velocity. Changes occur, according to the formula for the shear modulus of muscle. Organize the elastic modulus of each spatial point into a continuous field function. , representing the distribution of the elastic modulus at each point, where The spatial coordinates are superimposed on the patient's facial skull and tissue scan model;

[0041] In S2, the specific method for inferring the fat slip path is as follows:

[0042] S2-1, After removing the portion of the mandible that needs to be shaved from the facial skull and tissue scanning model, In the detection space, coordinate points where abrupt changes occur are identified to determine the modulus gradient. Greater than the mutation threshold The point, i.e., the elastic mutation point. ,in , represents the vector synthesized from the partial derivative arrays in the three directions;

[0043] S2-2, Let the response time of the shear wave in the three-dimensional tissue image be... , Indicates the shear wave at point At the time of the first arrival, the propagation direction vector of the shear wave at each point is defined as: That is, along the direction where the propagation time decreases, where After normalization, the unit propagation direction is obtained. , indicating the unit direction of propagation of the shear wave at this point;

[0044] S2-3, at a certain mutation point At that point, trace the propagation direction of the shear wave. And extend to generate trajectory lines ,in For path length, The size of the mutation points in images of fat distribution in patients at rest and under specific facial expressions. Distance between the centroids of the fat lobes Proportional, that is, at a certain mutation point The more easily the fat in a certain area is pulled by the covering muscle tissue, the greater the distance the fat will slide after jaw reduction surgery.

[0045] First, the shear elastic modulus of the muscles in the patient's mandible is measured. Based on the measurement results, the muscle mutation before and after mandibular bone removal is simulated. The mutation points and directions of the muscles are analyzed, thereby simulating the movement trajectory of the deep fat under the patient's muscles. Compared with directly generating postoperative fat distribution images, this simulation result has sufficient realistic basis for the displacement before and after surgery, and can more accurately simulate the displacement effect of fat.

[0046] In S3, the method for determining the complexity of muscle layers in the mandibular region is as follows: the muscles covering deep fat are divided according to the location of each fat lobe, and the muscle elasticity complexity index is used. ,in This is the point of abrupt change in the elastic modulus. This represents the total number of muscle mutation points located at the site of the fat lobe. A higher number of mutations and a greater degree of change in the direction of shear wave propagation indicate more complex muscle tissue. The muscle layers are highly complex at that time. To determine the complex threshold, the inferred fat slip path needs to be introduced into the interpolation algorithm;

[0047] In S3, the optimization and contour smoothing of easily foldable regions in the predicted image specifically involves: based on the predicted fat slip path... The algorithm generates a fat distribution structure in the mandibular region, and simultaneously uses an interpolation algorithm to generate a fat distribution structure based on the preoperative fat distribution structure. This interpolation is performed when a certain coordinate point in the preoperative fat distribution structure is located within the adipose tissue. To the nearest slip path closest distance At that time, the interpolation weight function ,in For the set radius of influence, when hour, The weight of a point in the interpolation is determined by its distance from the sliding path. The size of a certain coordinate point The closer it is to 1, the more it tends to use the fat sliding path. The resulting fat distribution structure in the mandibular region The closer it is to 0, the more it tends to the fat distribution structure generated by the difference algorithm.

[0048] This invention utilizes MRI fat-water separation imaging to detect the fat distribution structure in the mandibular region. Then, it uses ultrasound shear wave elastography combined with facial expression dynamic response to detect elastic mutations and propagation shifts in the muscle region. Based on this, it constructs an inverse modeling relationship between muscle and fat coupling to indirectly infer the fat slippage path. In areas with complex deep fat layers in the mandible, the image is reconstructed based on the inferred fat slippage path. For other locations, an interpolation algorithm is used to construct the contour, thereby correcting areas with tissue misalignment and image folding problems in image modeling. This prevents image distortion even when the patient's mandibular region has complex layers.

[0049] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0050] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A visual facial cosmetic surgery AI simulation method, characterized by: Includes the following steps: S1. First, visual technology and CT technology are used to obtain a scan model of the patient's facial skull and tissues. MRI water-fat separation imaging and ultrasound shear wave elastography are used to collect the fat distribution structure in the mandibular region using the MRI scanning unit. The static shear wave propagation is detected using an ultrasound probe and a high-speed receiver, and muscle elastic modulus map and fat distribution image are output. S2. Remove the part of the mandible that needs to be removed from the facial skull and tissue scanning model, extract the shear wave propagation direction and elastic mutation point, and construct a muscle and fat coupling model by combining image structure and elastic response. Based on the changes in fat distribution images of the patient's face under specific expression states, indirectly infer the fat slippage path before and after the mandibular bone removal surgery. S3. Based on the complexity of the muscle layers above the deep fat in the mandibular region, the predicted slip path is introduced into the tissue deformation simulation to identify and correct misaligned and folded areas. S4. Use the AI ​​image reconstruction module to generate a postoperative simulation 3D visualization image to obtain the correct fat distribution structure in the mandibular region. In S2, the specific method for inferring the fat slippage path is as follows: S2-1, After removing the portion of the mandible that needs to be shaved from the facial skull and tissue scanning model, In the detection space, coordinate points where abrupt changes occur are identified to determine the modulus gradient. Greater than the mutation threshold The point, i.e., the elastic mutation point. ,in ; S2-2, Let the response time of the shear wave in the three-dimensional tissue image be... , Indicates the shear wave at point At the time of the first arrival, the propagation direction vector of the shear wave at each point is defined as: That is, along the direction where the propagation time decreases, where After normalization, the unit propagation direction is obtained. , indicating the unit direction of propagation of the shear wave at this point; S2-3, at a certain mutation point At that point, trace the propagation direction of the shear wave. And extend to generate trajectory lines ,in For path length, The size of the mutation points in images of fat distribution in patients at rest and under specific facial expressions. Distance between the centroids of the fat lobes It is directly proportional.

2. The AI ​​simulation method for visual facial cosmetic surgery according to claim 1, characterized in that: In S1, the specific method for determining the fat distribution structure in the mandibular region is as follows: S1-1, Start the MRI scanning unit to scan the patient's face. The time-domain signal intensity of the MRI signal is: ,in For water signal strength, For fat signal intensity, The angular frequency of the water signal. The angular frequency of the fat signal. To represent the rotational change of the signal as a complex exponential function, and since the resonant frequency of fat is lower than that of water, the in-phase time point is selected. and the inverse time point This causes the water and fat signals to be in phase and out of phase at these two time points, respectively, based on the brightness of the in-phase image. Brightness of inverted image Find the sum and The size of the image is used to obtain the brightness of the water tissue image and the fat tissue image. When the value exceeds the preset value, it is determined to be adipose tissue. The water-like tissue containing muscle and soft tissue and the adipose tissue containing fat are distinguished to obtain the fat distribution structure in the mandibular region. The structure is fitted onto the facial skull and tissue scanning model and divided into multiple fat lobes according to the contour features.

3. The AI ​​simulation method for visualizing facial cosmetic surgery according to claim 2, characterized in that: In S1, the method for obtaining the muscle elastic modulus map is as follows: S1-2. Using an ultrasound probe, a scan is performed above the adipose tissue area in the mandibular region, which is identified as deep fat. A high-intensity, short-duration focused ultrasound shock wave is emitted, generating a localized, minute lateral displacement within the muscle covering the adipose tissue, producing a shear wave. This shear wave propagates outward from the excitation point, passing through tissues of varying elasticity, thus increasing the shear wave velocity. Changes occur, according to the formula for the shear modulus of muscle. Organize the elastic modulus of each spatial point into a continuous field function. , representing the distribution of the elastic modulus at each point, where The coordinates are superimposed on the patient's facial skull and tissue scan model.

4. The AI ​​simulation method for visual facial cosmetic surgery according to claim 3, characterized in that: In S3, the method for determining the complexity of muscle layers in the mandibular region is as follows: the muscles covering deep fat are divided according to the location of each fat lobe, and the muscle elasticity complexity index is used. ,in This is the point of abrupt change in elastic modulus. The total number of muscle mutation points at the location of the fat lobe, when The muscle layers are highly complex at that time. This is a threshold for complex judgments.

5. The AI ​​simulation method for visual facial cosmetic surgery according to claim 4, characterized in that: In step S3, optimizing and smoothing the contours of easily foldable regions in the predicted image specifically involves: based on the predicted fat slip path... The algorithm generates a fat distribution structure in the mandibular region, and simultaneously uses an interpolation algorithm to generate a fat distribution structure based on the preoperative fat distribution structure. When the preoperative fat distribution structure is located at a certain coordinate point in the adipose tissue... To the nearest slip path closest distance At that time, the interpolation weight function ,in For the set radius of influence, when hour, The weight of a point in the interpolation is determined by its distance from the sliding path. The size of a certain coordinate point The closer it is to 1, the more it tends to use the fat sliding path. The resulting fat distribution structure in the mandibular region The closer it is to 0, the more it tends to the fat distribution structure generated by the difference algorithm.

6. The visualized facial cosmetic surgery AI simulation system according to claim 5, characterized in that: The system includes an image acquisition module, a dynamic modeling module, and an image reconstruction module. The image acquisition module, based on MRI water-lipid separation imaging technology, acquires and identifies the three-dimensional distribution of the bone, muscle, fat, and epidermal structures of the patient's face, and constructs local elastic modulus maps of muscles under static and specific facial expression states based on ultrasound shear wave elastography. The dynamic modeling module is used to combine the detected distribution structure of the mandibular region with the elastic modulus abrupt changes and propagation direction of the muscle region to analyze the coupling behavior of deep fat and muscle and indirectly infer the fat slippage path. The image reconstruction module is used to construct a postoperative tissue slippage model, correct areas where tissue misalignment and image folding problems occur in advance during image modeling, and generate a postoperative predictive three-dimensional image.

7. The visualized facial cosmetic surgery AI simulation system according to claim 6, characterized in that: The image acquisition module includes an ultrasound probe, a high-speed ultrasound receiving unit, a shear wave velocity calculation module, an elastic modulus mapping module, an MRI scanning unit, and a water-lipid separation image module. The ultrasound probe excites transverse shear waves within the tissue using short pulses. The high-speed ultrasound receiving unit captures the signal of the shear wave propagating along the tissue. The shear wave velocity calculation module calculates the propagation velocity of the wavefront at different depths and directions. The elastic modulus mapping module calculates the regional elastic modulus and generates an elastic distribution map. The MRI scanning unit scans the face and constructs a preoperative three-dimensional spatial model. The water-lipid separation image module performs water-lipid separation image reconstruction on the acquired multi-phase data to generate a distribution structure map of the mandibular region. The dynamic modeling module includes an elastic mutation extraction module, a shear wave direction analysis module, a coupling modeling module, and a slip path generation module. The elastic mutation extraction module is used to analyze the spatial mutation points of the elastic modulus in the shear wave imaging image to infer the initiation source of slip. The shear wave direction analysis module is used to track the propagation path of the shear wave and infer the potential direction of fat movement. The coupling modeling module is used to construct a coupling model based on the elastic response of muscle and the static distribution of fat. The slip path generation module is used to infer the slip path of fat. The image reconstruction module includes a complexity judgment module, a smooth transition processing module, an interpolation algorithm optimization module, and an AI facial reconstruction module. The complexity judgment module is used to determine the hierarchical complexity of each region in the patient's mandibular region. The smooth transition processing module is used to stretch the boundary surface to make the contour look coherent in order to handle the problems of folds and grooves after facial image reconstruction. The interpolation algorithm optimization module is used to optimize the interpolation algorithm for smooth transition processing after the interpolation algorithm provides the inferred fat slip path. The AI ​​facial reconstruction module is used to reconstruct the three-dimensional image of the skull after the patient's mandibular osteotomy using AI prediction methods.

Images