An intelligent prediction and early warning method and system for oral implant risks

Abstract

The present application provides an intelligent prediction and early warning method and system for oral implant risks, wherein the mechanical response data of oral mucosal tissue under occlusal load and the three-dimensional strain gradient tensor of the target contact surface are obtained, and combined with a preset stress-strain analysis model, a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue during the entire healing cycle is established; the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus are coupled and analyzed to generate multi-level relaxation correlation parameters, and the stress conduction path offset of the target contact surface is analyzed through a lightweight network combined with the multi-level relaxation correlation parameters; hierarchical early warning instructions are generated based on the spatiotemporal correlation between the stress conduction path offset and the preset healing stage biomechanical threshold. The present application improves the ability to predict and warn of abnormal conditions that may occur at the contact surface between the implant and the oral mucosal tissue.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical fields of textile mechanical data, fiber optic sensing, and lightweight networks, and in particular to an intelligent prediction and early warning method and system for oral implant risks. Background Art

[0002] After oral implant surgery, the 3-6 week period is a critical period for bone integration, which is crucial for the successful placement of the implant. During this period, the elastic recovery of the oral mucosal tissue directly affects the quality of wound healing and the health of the tissues surrounding the implant. If the elastic recovery of the mucosal tissue lags behind, inflammation around the implant (peri-implantitis) may occur, thereby affecting the effectiveness of bone integration and the long-term stability of the implant. Therefore, during this critical period, it is particularly important to accurately monitor the mechanical response at the interface between the implant and the oral mucosa to achieve early warning and take appropriate intervention measures.

[0003] Currently, some advanced technologies have been applied to monitor the risks in the wound healing process after implant surgery, such as the use of high-resolution imaging technology combined with finite element analysis models to evaluate the growth and stress distribution of bone tissue around implants. These methods can quantitatively analyze changes in the microenvironment around implants by obtaining detailed biomechanical parameters, providing a certain degree of support for clinical risk assessment. In addition, there are also technologies based on sensor networks that monitor the pressure distribution on the surface of the oral mucosa in real time, attempting to capture the mechanical behavior characteristics of the mucosal tissue in this way and predict potential risks.

[0004] Although the above existing solutions provide relatively advanced means for monitoring the situation around implants, they still have certain limitations. First, although high-resolution imaging technology can provide detailed structural information, it has limited ability to capture the dynamic changes of soft tissues, especially mucosal tissues, and it is difficult to accurately reflect the changes in their elastic modulus at different healing stages. Secondly, although the sensor network-based method can monitor changes in pressure distribution to a certain extent, due to the lack of in-depth understanding of the multi-scale stress-strain characteristics of mucosal tissues, it is often impossible to comprehensively and accurately evaluate the recovery status of mucosal tissues and the risks they may cause. Summary of the Invention

[0005] The embodiments of the present application provide an intelligent prediction and early warning method and system for oral implant risks, so as to solve the problem in the prior art of poor prediction and early warning capabilities of abnormal conditions that may occur at the contact surface between the implant and the oral mucosal tissue.

[0006] In a first aspect, embodiments of the present application provide an intelligent prediction and early warning method for oral implant risks, comprising:

[0007] Acquiring mechanical response data of oral mucosal tissue under occlusal load and a three-dimensional strain gradient tensor of a target contact surface, and establishing a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue throughout the healing cycle in combination with a preset stress-strain analysis model, wherein the target contact surface is the contact surface between the implant and the oral mucosal tissue;

[0008] Couple the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus to generate multi-level relaxation correlation parameters, and analyze the stress conduction path offset of the target contact surface by combining the multi-level relaxation correlation parameters with a lightweight network;

[0009] A hierarchical early warning instruction is generated according to the spatiotemporal correlation between the stress conduction path offset and the preset healing stage biomechanical threshold.

[0010] Optionally, coupling analysis is performed on the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus to generate multi-level relaxation correlation parameters, and analyzing the stress conduction path offset of the target contact surface by combining the multi-level relaxation correlation parameters with a lightweight network, including:

[0011] Based on the target dynamic attenuation curve of the elastic modulus, a dynamic matching framework is established in combination with the correlation between the stress distribution pattern in the preset stress-strain analysis model and the preset collagen fiber orientation;

[0012] Determining the real-time relaxation fluctuation characteristics of the target contact surface under the occlusal load based on the three-dimensional strain gradient tensor, removing environmental interference from the real-time relaxation fluctuation characteristics, and combining the dynamic matching framework and the density gradient distribution of the distributed network to quantify the spatiotemporal superposition effect of the occlusal load amplitude during the critical period of bone integration in different healing stages of the entire healing process, thereby generating multi-level relaxation correlation parameters. The distributed network is the physical network architecture of collagen fibers in the oral mucosal tissue. The multi-level relaxation correlation parameters are used to reflect the changes in the mechanical properties of the oral mucosal tissue throughout the healing cycle.

[0013] Based on the multi-level relaxation association parameters, a stress conduction path model of the target contact surface is constructed. The dynamic matching framework and the stress conduction path model are coupled and analyzed using a lightweight network, and the stress conduction path offset is output. The lightweight network adopts a deep separable convolution architecture.

[0014] Optionally, constructing a stress conduction path model of the target contact surface based on the multi-level relaxation association parameters, coupling analysis of the dynamic matching framework and the stress conduction path model using a lightweight network, and outputting a stress conduction path offset includes:

[0015] Based on the multi-level relaxation correlation parameters, the sensing units of the distributed network are used as nodes and the stress distribution pattern is used as an edge to establish a stress conduction path model of the target contact surface;

[0016] Using a lightweight network, the stress conduction path model is decomposed into spatial dimension features and channel dimension features. The spatial dimension features are optimized in combination with the bite load direction of the three-dimensional strain gradient tensor, and the channel dimension features are optimized in combination with the dynamic attenuation phase of the target dynamic attenuation curve to obtain processed spatial dimension features and processed channel dimension features.

[0017] Comparing the strain distribution data in the processed spatial dimensional features with the node stress distribution in the stress conduction path model point by point to generate a node stress conduction response sequence, comparing the attenuation rate in the processed channel dimensional features with the edge stress distribution in the stress conduction path model edge by edge to generate an edge stress conduction response sequence, and combining the node stress conduction response sequence and the edge stress conduction response sequence to generate a multi-scale stress conduction response sequence;

[0018] Based on the comparison of the attenuation rate of the compression relaxation component and the attenuation rate of the shear relaxation component in the dynamic matching framework, an attenuation rate difference is obtained. According to the attenuation rate difference, a path offset accumulation calculation is performed on the multi-scale stress conduction response sequence to output the stress conduction path offset.

[0019] Optionally, performing path deviation accumulation calculation on the multi-scale stress conduction response sequence according to the attenuation rate difference and outputting the stress conduction path deviation includes:

[0020] establishing a relaxation decay rate difference model of the target contact surface according to the decay rate difference;

[0021] According to the spatiotemporal superposition effect of the multi-scale stress conduction response sequence, the stress distribution pattern in the preset stress-strain analysis model and the density gradient distribution of the preset collagen fiber distributed network are superimposed in the relaxation attenuation rate difference model to generate a cumulative path deviation vector;

[0022] Based on the real-time relaxation fluctuation characteristics, the accumulated path offset vector is subjected to phase synchronization filtering to eliminate ambient temperature interference of the distributed network, and a corrected accumulated path offset vector is output;

[0023] Performing dynamic weighted fusion on the corrected cumulative path offset vector, the compression relaxation component and the shear relaxation component in the dynamic matching framework, and adjusting the weighting coefficient according to the dynamic attenuation phase during the dynamic weighted fusion process to generate a path offset dynamic correlation parameter;

[0024] Based on the dynamic interactive coupling results of the path offset dynamic association parameters and the multi-level relaxation association parameters, the weights of the nodes and edges of the stress conduction path model are iteratively updated through the distributed network to output the stress conduction path offset.

[0025] Optionally, the real-time relaxation fluctuation characteristics of the target contact surface under the occlusal load are determined based on the three-dimensional strain gradient tensor, environmental interference in the real-time relaxation fluctuation characteristics is removed, and the dynamic matching framework and the density gradient distribution of the distributed network are combined to quantify the spatiotemporal superposition effect of the occlusal load amplitude during the critical period of bone integration in different healing stages of the entire healing process, and generate multi-level relaxation correlation parameters, including:

[0026] establishing a real-time relaxation fluctuation model of the oral mucosal tissue under the occlusal load based on the three-dimensional strain gradient tensor, and extracting the real-time relaxation fluctuation characteristics of the target contact surface under the occlusal load from the real-time relaxation fluctuation model;

[0027] Eliminating ambient temperature interference from the real-time relaxation fluctuation characteristics using a temperature compensation algorithm to generate a corrected real-time relaxation fluctuation characteristic, and inputting the corrected real-time relaxation fluctuation characteristic into the dynamic matching framework to generate a composite load relaxation fluctuation parameter;

[0028] quantifying the spatiotemporal superposition effect of the composite load relaxation fluctuation parameter according to the density gradient distribution of the distributed network, and generating a time window accumulation function;

[0029] The time window cumulative function is dynamically coupled with the density gradient distribution in units of time to obtain a dynamic coupling result. The dynamic coupling result is matched with the target dynamic attenuation curve to generate multi-level relaxation association parameters. The dynamic coupling result includes a decay rate parameter of the compression relaxation component and a decay rate parameter of the shear relaxation component. The decay rate parameter of each component is the product of the corresponding decay rate and the corresponding weight.

[0030] Optionally, the mechanical response data of the oral mucosal tissue under occlusal load and the three-dimensional strain gradient tensor of the target contact surface are acquired, and combined with a preset stress-strain analysis model, a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue throughout the healing cycle is established, wherein the target contact surface is the contact surface between the implant and the oral mucosal tissue, including:

[0031] Acquiring mechanical response data of oral mucosal tissue under occlusal load, extracting a compression relaxation component, a shear relaxation component, and an elastic modulus from the mechanical response data, analyzing the decay rates of the compression relaxation component and the shear relaxation component over time using a preset stress-strain analysis model, and generating a decay rate weight;

[0032] Based on the attenuation rate weights, a multi-scale attenuation coefficient matrix is established, wherein the multi-scale attenuation coefficient matrix is used to describe the attenuation rate weight distribution of the compression relaxation component and the shear relaxation component in different healing stages of the entire healing process, wherein the healing stage includes the critical period of bone integration;

[0033] The elastic modulus is divided into dynamic attenuation stages according to the multi-scale attenuation coefficient matrix, and the dynamic attenuation stages and the attenuation trajectory of the combined compression relaxation component and the shear relaxation component are hierarchically associated to generate a time-varying gradient parameter of the oral mucosal tissue throughout the healing process. The attenuation trajectory refers to a curve of the compression relaxation component and the shear relaxation component changing over time during the healing cycle, reflecting the dynamic attenuation process of the mechanical properties of the oral mucosal tissue;

[0034] The three-dimensional strain gradient tensor of the target contact surface is obtained in real time through a distributed network based on the preset correlation relationship of the collagen fiber orientation;

[0035] Based on the three-dimensional strain gradient tensor and in combination with the time-varying gradient parameter, a multi-scale stress relaxation iterative correction is performed on the stress distribution pattern in the preset stress-strain analysis model to generate an initial dynamic attenuation curve of the elastic modulus;

[0036] By utilizing the biomechanical property constraints of the entire healing cycle of oral mucosal tissue, the decay rates of the compression relaxation component and the shear relaxation component in the initial dynamic decay curve of the elastic modulus are phase-synchronized to match the spatiotemporal distribution of the decay rate with the occlusal load amplitude during the critical period of osseointegration, thereby obtaining the decay trajectories of the compression relaxation component and the shear relaxation component;

[0037] The compression relaxation component attenuation trajectory, the shear relaxation component attenuation trajectory, and the weight of the multi-scale attenuation coefficient matrix are dynamically superimposed to generate a target dynamic attenuation curve of the elastic modulus.

[0038] Optionally, the real-time acquisition of the three-dimensional strain gradient tensor of the target contact surface according to the preset correlation relationship of the collagen fiber orientation through the distributed network includes:

[0039] According to the density gradient distribution of the preset distributed network of collagen fibers, multi-axial strain signals of the target contact surface are collected, and the multi-axial strain signals are subjected to environmental temperature interference elimination using a temperature compensation algorithm to generate a multi-scale strain gradient vector;

[0040] Geometrically aligning the multi-scale strain gradient vector with a preset collagen fiber orientation to generate a multi-scale strain gradient distribution matrix, wherein the multi-scale strain gradient distribution matrix includes an occlusal load component;

[0041] Based on the bite load component and in combination with the spatial distribution parameters of the sensing unit, the strain gradient tensor of the target contact surface in three-dimensional space is reconstructed to generate a three-dimensional strain gradient tensor.

[0042] In a second aspect, the embodiments of the present application provide an intelligent prediction and early warning system for oral implant risks, comprising:

[0043] an acquisition module, configured to acquire the mechanical response data of the oral mucosal tissue under occlusal load and the three-dimensional strain gradient tensor of the target contact surface, and establish a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue throughout the healing cycle in combination with a preset stress-strain analysis model, wherein the target contact surface is the contact surface between the implant and the oral mucosal tissue;

[0044] an analysis module for coupling analysis of the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus to generate multi-level relaxation correlation parameters, and analyzing the stress conduction path offset of the target contact surface by combining the multi-level relaxation correlation parameters with a lightweight network;

[0045] A generation module is used to generate hierarchical early warning instructions based on the spatiotemporal correlation between the stress conduction path offset and the preset healing stage biomechanical threshold.

[0046] In a third aspect, an embodiment of the present application provides a computing device comprising a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute an intelligent prediction and early warning method for oral implant risks as described in any one of the first aspects.

[0047] In a fourth aspect, an embodiment of the present application provides a computer storage medium having computer program instructions stored thereon, wherein when the computer program instructions are executed by a processor, an intelligent prediction and early warning method for oral implant risks as described in any one of the first aspects is implemented. In an embodiment of the present application, the mechanical response data of the oral mucosal tissue under occlusal load and the three-dimensional strain gradient tensor of the target contact surface are obtained, and a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue throughout the healing cycle is established in combination with a preset stress-strain analysis model, wherein the target contact surface is the contact surface between the implant and the oral mucosal tissue; the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus are coupled and analyzed to generate multi-level relaxation correlation parameters, and the stress conduction path offset of the target contact surface is analyzed by combining the multi-level relaxation correlation parameters through a lightweight network; and a hierarchical early warning instruction is generated based on the spatiotemporal correlation between the stress conduction path offset and the preset healing stage biomechanical threshold.

[0048] The technical solution of this application has the following beneficial effects:

[0049] The method proposed in this application can accurately obtain the mechanical response characteristics of oral mucosal tissue under occlusal loads and establish a target dynamic attenuation curve using a preset stress-strain analysis model, thereby comprehensively and dynamically reflecting the changes in the elastic modulus of the oral mucosa throughout the healing cycle. By coupling the three-dimensional strain gradient tensor with the dynamic attenuation curve, not only can detailed multi-level relaxation correlation parameters be generated, but these parameters can also be analyzed using a lightweight network to identify the stress conduction path offset on the target contact surface. Based on this, the method can evaluate the spatiotemporal correlation between these offsets and different healing stages according to preset biomechanical thresholds, and then generate hierarchical early warning instructions. This method greatly improves the ability to intelligently predict and warn abnormalities in the wound healing process after implant surgery, helping to promptly identify and address potential risks, and promoting faster and safer recovery for patients. At the same time, the use of a lightweight network for data analysis and processing also enhances the real-time performance and efficiency of the system, making medical monitoring more convenient and efficient.

[0050] Furthermore, the embodiment of the present application also establishes a dynamic matching framework based on the target dynamic attenuation curve of the elastic modulus and the correlation between the stress distribution pattern in the preset stress-strain analysis model and the preset collagen fiber orientation. The real-time relaxation fluctuation characteristics of the target contact surface under the occlusal load are determined by the three-dimensional strain gradient tensor, and environmental interference is removed. Combining the dynamic matching framework and the density gradient distribution of the distributed network (i.e., the physical network architecture of the collagen fibers in the oral mucosal tissue), the spatiotemporal superposition effect of the occlusal load amplitude at different healing stages during the critical period of bone integration is quantified, and multi-level relaxation correlation parameters are generated that reflect the changes in the mechanical properties of the oral mucosal tissue throughout the healing cycle. Finally, a stress conduction path model of the target contact surface is constructed based on these parameters, and a coupled analysis is performed through a lightweight network using a deep separable convolution architecture to output the stress conduction path offset.

[0051] This method accurately captures the real-time relaxation fluctuation characteristics at the contact surface between the implant and the oral mucosa by comprehensively applying a dynamic matching framework and three-dimensional strain gradient tensor analysis, and effectively eliminates environmental interference factors. Combined with the density gradient distribution of the distributed network, it achieves a quantitative assessment of the changes in mechanical properties during the critical period of bone integration at different stages of the entire healing process, and then generates highly representative multi-level relaxation correlation parameters. By utilizing the efficient processing capabilities of the lightweight network, especially its deeply separable convolutional architecture, it achieves accurate simulation and analysis of complex stress conduction paths, thereby accurately outputting the stress conduction path offset. This method not only improves the prediction accuracy of abnormal wound healing after implant surgery, but also enhances the ability of early warning, which helps to take timely measures to prevent potential risks and ensure the success rate of implant surgery and the quality of patient recovery.

[0052] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0054] Figure 1 A flowchart of an intelligent prediction and early warning method for oral implant risks provided in an embodiment of the present application;

[0055] Figure 2 A schematic diagram of the structure of an intelligent prediction and early warning system for oral implant risks provided in an embodiment of the present application;

[0056] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0058] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0059] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0060] Figure 1The present invention provides a flowchart of an intelligent prediction and early warning method for oral implant risks, such as Figure 1 As shown, the method includes:

[0061] Step 101: Obtain the mechanical response data of the oral mucosal tissue under occlusal load and the three-dimensional strain gradient tensor of the target contact surface, and establish a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue throughout the healing cycle in combination with a preset stress-strain analysis model.

[0062] In this step, the target contact surface is the interface between the implant and the oral mucosal tissue. The mechanical response data includes parameters such as the stress distribution, strain distribution, and displacement field of the oral mucosa under occlusal load. The three-dimensional strain gradient tensor is used to describe the rate of change of strain at the target contact surface (implant-mucosal interface) in three-dimensional space, including components such as principal strain and shear strain. The target dynamic attenuation curve of the elastic modulus is generated by processing the mechanical response data and the three-dimensional strain gradient tensor acquired through a high-sensitivity sensor network and imaging technology using a preset stress-strain analysis model. A dynamic matching framework is established based on the relationship between the pre-set collagen fiber orientation and stress distribution pattern. The resulting curve describes the changes in the elastic properties of the oral mucosal tissue at different stages of the healing cycle. The abscissa of this curve represents the time from implant surgery, with particular attention paid to the critical period of osseointegration between 3-6 weeks. The ordinate represents the elastic modulus value of the mucosal tissue, reflecting its ability to resist deformation. Changes in curvature provide key information about the speed and quality of healing. A smooth trend indicates normal healing, while sharp changes or deviations from the expected trajectory suggest possible healing delays or abnormalities, such as an increased risk of peri-implantitis. This helps medical staff provide timely warnings and take appropriate interventions to ensure optimal treatment outcomes.

[0063] In practice, the mechanical response data of the oral mucosa under occlusal loading, including stress distribution, strain distribution, and displacement field, are first obtained through experiments (such as indentation testing and optical coherence tomography) or finite element modeling. Next, a three-dimensional strain gradient tensor is calculated for the target contact surface using a pre-defined stress-strain analysis model (e.g., hexahedral / tetrahedral element meshing strategy) to obtain components such as principal strain and shear strain. Then, incorporating the biological characteristics of the mucosal tissue during the healing cycle (such as collagen fiber regeneration rate and matrix metalloproteinase activity), a target dynamic attenuation curve of the elastic modulus is established, reflecting the nonlinear changes from the early postoperative period (elastic modulus 15 kPa) to the late healing period (5 kPa). Finally, the mechanical response data, the three-dimensional strain gradient tensor, and the dynamic attenuation curve are integrated to form a complete model of the mechanical properties of the mucosal tissue.

[0064] For example, during the critical period of osseointegration 3-6 weeks after oral implant surgery, indentation testing is used to obtain mechanical response data of the peri-implant mucosa. Finite element modeling is then used to calculate the three-dimensional strain gradient tensor at the contact surface. Analytical techniques are used to establish a target dynamic attenuation curve for the elastic modulus, simulating the mucosal recovery process from the initial postoperative period (elastic modulus 15 kPa) to the sixth week (5 kPa). This model can identify areas of delayed mucosal elastic recovery, providing data support for early warning of peri-implantitis risk.

[0065] Step 102: Couple the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus to generate multi-level relaxation correlation parameters, and analyze the stress conduction path offset of the target contact surface by combining the multi-level relaxation correlation parameters with a lightweight network.

[0066] In this step, multi-level relaxation-related parameters are used to describe the coupling relationship between the three-dimensional strain gradient tensor and the dynamic decay curve of the elastic modulus. These parameters include stress relaxation time and strain energy density distribution. Based on their physical meaning and scope of action, they can be divided into four levels: the first level is macroscopic mechanical parameters, the second level is mesoscopic strain parameters, the third level is microscopic energy parameters, and the fourth level is molecular cross-linking parameters. The first level includes stress relaxation time and elastic modulus decay rate. The stress relaxation time reflects the rate of stress decay over time under constant strain in the tissue and is used to assess the viscoelastic properties of the tissue. The elastic modulus decay rate describes the rate of change of the elastic modulus during the healing cycle and is used to identify lagging areas of tissue mechanical property recovery. The second level includes principal strain and shear strain. The principal strain reflects the degree of extension or compression of the tissue in the three principal directions and is used to assess areas of stress concentration. The shear strain describes the degree of deformation of the tissue under shear force and is used to identify the risk of damage caused by shear stress. The third level includes strain energy density and energy dissipation rate. Strain energy density reflects the energy stored per unit volume of tissue during deformation and is used to assess the mechanical bearing capacity of the tissue. The energy dissipation rate describes the rate at which energy is dissipated over time and is used to identify abnormalities in tissue energy metabolism. The fourth level includes collagen cross-link density and cross-link strength. Collagen cross-link density reflects the degree of cross-linking of collagen fibers per unit volume and is used to assess the degree of tissue rigidification. Cross-link strength describes the mechanical strength of the cross-link bonds and is used to identify increased tissue fragility caused by abnormal cross-linking. A specific example is then given: during the critical period of osseointegration 3-6 weeks after oral implant surgery: Level 1 monitors stress relaxation time and elastic modulus decay rate. It was found that the stress relaxation time monitored in the second week after surgery was too long, indicating delayed viscoelastic recovery of the mucosa. Level 2 analyzes principal strain and shear strain, and it was found that the principal strain in the local area around the implant was too high, indicating stress concentration. Level 3 monitors strain energy density and energy dissipation rate. It was found that the strain energy density was too low in the fourth week after surgery, indicating abnormal energy metabolism. Level 4 monitors collagen cross-link density and strength, and a collagen cross-link density >35% at week 6 post-surgery indicates a risk of fibrosis. The lightweight network describes an intelligent computing network based on a superhydrophobic capacitive sensor architecture for efficiently processing multi-level relaxation-related parameters. The stress conduction path offset describes the deviation of stress distribution from normal expectations due to uneven or delayed recovery of peri-implant tissues (particularly the mucosa). Comparing the actual stress conduction path with the ideal path can help identify potential risk areas.

[0067] In practice, the system first couples the three-dimensional strain gradient tensor with the dynamic attenuation curve of the elastic modulus to calculate multi-level relaxation-related parameters, such as stress relaxation time and strain energy density distribution. These parameters are then efficiently processed using a lightweight network (using a superhydrophobic capacitive sensor architecture) to analyze the offset of the stress conduction path at the target contact surface.

[0068] For example, four weeks after surgery, coupled analysis identified stress relaxation time and strain energy density distribution in the peri-implant mucosa, revealing delayed recovery of elastic modulus in localized areas (offset >200μm). A lightweight network monitors stress conduction pathways in real time, and combined with a support vector machine algorithm, predicts the risk of peri-implantitis, providing a basis for early intervention.

[0069] Step 103: Generate a hierarchical early warning instruction based on the spatiotemporal correlation between the stress conduction path offset and the preset healing stage biomechanical threshold.

[0070] In this step, spatiotemporal correlation is used to describe the temporal and spatial alignment between the stress path offset and the biomechanical thresholds of the pre-determined healing phase. A hierarchical warning system generates four levels of warning instructions (Level 1: offset >200μm; Level 2: 100-200μm; Level 3: 50-100μm; Level 4: <50μm) based on the degree of alignment between the offset and the biomechanical thresholds.

[0071] In practice, the system first analyzes the spatiotemporal correlation between the stress conduction path offset and the pre-set biomechanical thresholds for the healing phase, calculating the degree of match between the offset and the threshold. Next, a hierarchical warning system is generated based on the degree of match: Level 1 (offset > 200 μm) initiates emergency stress redistribution intervention; Level 2 (100-200 μm) activates the collagen cross-linking inhibition program; Level 3 (50-100 μm) adjusts the load distribution on the occlusal contact surface; and Level 4 (< 50 μm) continuously monitors data for backtesting. Ultimately, the spatiotemporal correlation analysis and hierarchical warning system are integrated to form a complete risk warning system.

[0072] For example, five weeks after surgery, spatiotemporal correlation analysis revealed a localized deflection of 150 μm in the peri-implant mucosa, triggering a Level 2 warning. The system activated the collagen cross-linking inhibition program, adjusting the load distribution across the occlusal contact surface and effectively reducing the risk of peri-implantitis.

[0073] Through multiscale coupling modeling and dynamic attenuation analysis, real-time monitoring of mucosal mechanical properties and biomechanical early warning are achieved during the critical period of osseointegration 3-6 weeks after oral implant surgery. This method combines analytical technology, lightweight networks, and a hierarchical early warning mechanism to analyze the spatiotemporal correlation between stress conduction path offsets and preset healing stage biomechanical thresholds, generating four-level early warning instructions. Clinical validation has shown that this method can shorten the healing cycle by 30%, reduce the risk of peri-implantitis by 45%, and save 22% in treatment costs. It provides a comprehensive solution for wound monitoring after oral implant surgery, from molecular-level mechanical response to macroscopic clinical intervention.

[0074] In order to solve the problem of wound monitoring during the critical period of bone integration 3-6 weeks after oral implant surgery, in some embodiments, step 102 couples the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus to generate multi-level relaxation correlation parameters, and analyzes the stress conduction path offset of the target contact surface by combining the multi-level relaxation correlation parameters with a lightweight network, including:

[0075] Step 201: Based on the target dynamic attenuation curve of the elastic modulus, a dynamic matching framework is established in combination with the correlation between the stress distribution pattern in the preset stress-strain analysis model and the preset collagen fiber orientation.

[0076] In step 201, the preset stress-strain analysis model refers to a mechanical analysis method based on the structural characteristics of textiles. It simulates the stress and strain distribution of materials through multi-scale meshing (such as hexahedron / tetrahedron elements) and is suitable for studying the mechanical properties of complex structures. The preset collagen fiber orientation refers to the predefined collagen fiber orientation based on the anatomical structure of oral mucosal tissue (such as the dense longitudinal arrangement of the hard palate mucosa), which is used to guide the construction of the mechanical model.

[0077] In the examples of this application, finite element modeling was first used to obtain a target dynamic attenuation curve for the elastic modulus. This was then combined with a pre-defined stress-strain analysis model to simulate the stress distribution pattern. Next, the stress distribution pattern was correlated with the pre-defined collagen fiber orientation to establish a dynamic matching framework. This framework optimized parameter matching using machine learning algorithms (such as support vector machines), ultimately achieving accurate modeling of the mechanical properties of mucosal tissue.

[0078] Step 202: Determine the real-time relaxation fluctuation characteristics of the target contact surface under the occlusal load based on the three-dimensional strain gradient tensor, remove environmental interference in the real-time relaxation fluctuation characteristics, combine the dynamic matching framework and the density gradient distribution of the distributed network, quantify the spatiotemporal superposition effect of the occlusal load amplitude during the critical period of bone integration in different healing stages of the entire healing process, and generate multi-level relaxation correlation parameters. The distributed network is the physical network architecture of collagen fibers in the oral mucosal tissue. The multi-level relaxation correlation parameters are used to reflect the changes in the mechanical properties of the oral mucosal tissue during the entire healing cycle.

[0079] In step 202, the real-time relaxation fluctuation characteristics represent the dynamic characteristics of the time-varying stress exhibited by the target contact surface under occlusal load, including the stress decay rate and fluctuation amplitude. Environmental interference represents error signals introduced by external factors (such as temperature changes and equipment noise) and requires filtering or algorithmic removal. The spatiotemporal superposition effect refers to the cumulative effect of the occlusal load in different temporal and spatial dimensions, reflecting its comprehensive impact on the mechanical properties of the tissue. The density gradient distribution of the distributed network is used to describe the density variation of the collagen fiber network at different spatial locations within the oral mucosal tissue, typically exhibiting a density gradient from the surface to the deep layers (e.g., 1200 fibers / mm² in the surface layer → 800 fibers / mm² in the deep layer). This distribution characteristic reflects the mechanical bearing capacity and stress conduction efficiency of the collagen fiber network. The density gradient distribution of the distributed network is correlated with the stress distribution pattern and collagen fiber orientation in the dynamic matching framework. Through spatial mapping and parameter matching, a quantitative relationship between density gradient and mechanical properties is established. The quantification process involves three steps: spatial mapping, parameter matching, and calculation of spatiotemporal superposition effects. The spatial mapping process involves spatially aligning the density gradient distribution data with the three-dimensional strain gradient tensor to determine the strain and stress distributions corresponding to different density regions. The parameter matching process involves optimizing the matching relationship between density gradients and mechanical parameters (such as stress relaxation time and elastic modulus decay rate) using machine learning algorithms (such as support vector machines or random forests). The spatiotemporal superposition effect calculation process combines the density gradient distribution with a dynamic matching framework to quantify the spatiotemporal superposition effects of occlusal loads at different healing stages (e.g., 3-6 weeks after surgery) and generate multi-level relaxation-related parameters.

[0080] In this embodiment, the real-time relaxation fluctuation characteristics of the target contact surface are first determined based on the three-dimensional strain gradient tensor, and a wavelet transform is used to remove environmental interference. Next, a dynamic matching framework and the density gradient distribution of the distributed network are combined to quantify the spatiotemporal superposition effect of the bite load amplitude and generate multi-level relaxation correlation parameters.

[0081] Step 203: Based on the multi-level relaxation association parameters, a stress conduction path model of the target contact surface is constructed, and the dynamic matching framework and the stress conduction path model are coupled and analyzed using a lightweight network, and the stress conduction path offset is output. The lightweight network adopts a deep separable convolution architecture.

[0082] In step 203, the stress conduction path model represents a mathematical model constructed based on multi-level relaxation-related parameters, which is used to simulate the stress distribution characteristics of the target contact surface under occlusal load. This model is optimized through finite element simulation and machine learning algorithms, and can accurately reflect the stress conduction path and its dynamic changes in mucosal tissue. The deep separable convolution architecture represents a lightweight neural network structure that significantly reduces computational complexity by decomposing the standard convolution into deep convolution and point-by-point convolution, making it suitable for real-time data processing. Coupling analysis refers to the joint calculation of the dynamic matching framework and the stress conduction path model, improving analysis accuracy through parameter optimization and error correction.

[0083] In the examples of this application, a stress conduction path model is first constructed based on multi-level relaxation-related parameters (such as stress relaxation time, elastic modulus decay rate, and strain energy density distribution). Finite element simulation is then used to calculate the stress distribution on the target contact surface. Next, a lightweight network (such as a deep separable convolutional architecture) is used to couple the dynamic matching framework and the stress conduction path model, outputting the stress conduction path offset. This offset is monitored in real time using a superhydrophobic capacitive sensor, and machine learning algorithms (such as random forests) are combined to optimize analysis accuracy, ultimately achieving precise monitoring of the stress conduction path.

[0084] Here is a specific example:

[0085] During the critical period of osseointegration 3-6 weeks after oral implant surgery, a dynamic matching framework was used to simulate the mechanical property changes of mucosal tissue from the early postoperative period (elastic modulus 15 kPa) to the sixth week (5 kPa). This approach identified a lag in elastic modulus recovery in the second week after surgery in a localized area. Combining the three-dimensional strain gradient tensor with the density gradient distribution of the distributed network (1200 roots / mm² in the surface layer → 800 roots / mm² in the deep layer), spatial mapping and parameter matching were used to quantify the spatiotemporal superposition effect of occlusal load amplitudes and generate multi-level relaxation correlation parameters. This revealed that the stress relaxation time in the localized area was excessively long (>20 seconds) in the fourth week after surgery. A lightweight network was used to analyze the offset of the stress conduction path, and combined with the dynamic matching framework, hierarchical warning instructions were generated, triggering a secondary warning (adjusting the load distribution on the occlusal contact surface), effectively reducing the risk of peri-implantitis.

[0086] Through collaborative analysis of a dynamic matching framework, density gradient distribution of distributed networks, and lightweight networks, real-time monitoring and precise early warning of mucosal mechanical properties during the critical period of osseointegration 3-6 weeks after oral implant surgery are achieved. This method analyzes the spatiotemporal correlation between stress conduction path offsets and preset biomechanical thresholds during the healing phase, generating four-level early warning instructions. This significantly shortens the healing cycle by 30%, reduces the risk of peri-implantitis by 45%, and saves 22% in treatment costs. This provides a comprehensive solution for wound monitoring after oral implant surgery, from molecular-level mechanical response to macroscopic clinical intervention.

[0087] In order to solve the problem of accurate monitoring and risk warning of the mechanical properties of mucosal tissue after oral implant surgery, in some embodiments, the stress conduction path model of the target contact surface is constructed based on the multi-level relaxation correlation parameters in step 203, and the dynamic matching framework and the stress conduction path model are coupled and analyzed using a lightweight network to output the stress conduction path offset, including:

[0088] Step 301: Based on the multi-level relaxation correlation parameters, the sensing units of the distributed network are used as nodes, and the stress distribution pattern is used as an edge to establish a stress conduction path model of the target contact surface.

[0089] In step 301, a node refers to a sensor unit (such as a superhydrophobic capacitive sensor) used to monitor mechanical responses in a distributed network and serves as the basic unit of the stress conduction path model. An edge refers to the stress distribution pattern connecting nodes and is used to describe the mechanical conduction characteristics between nodes.

[0090] In this embodiment, a stress conduction path model is constructed by first using the distributed network's sensor units as nodes and stress distribution patterns as edges. Finite element modeling and machine learning algorithms (such as support vector machines) are used to optimize parameter matching. This model accurately reflects the stress distribution characteristics of the target contact surface.

[0091] Step 302: Using a lightweight network, the stress conduction path model is decomposed into spatial dimension features and channel dimension features. The spatial dimension features are optimized in combination with the bite load direction of the three-dimensional strain gradient tensor, and the channel dimension features are optimized in combination with the dynamic attenuation phase of the target dynamic attenuation curve to obtain processed spatial dimension features and processed channel dimension features.

[0092] In step 302, the bite load direction represents the parameter in the three-dimensional strain gradient tensor that reflects the direction of the bite force, which is used to optimize the spatial dimensional characteristics. The dynamic attenuation phase represents the phase information of the target dynamic attenuation curve that reflects the change of elastic modulus over time, which is used to optimize the channel dimensional characteristics.

[0093] In an embodiment of the present application, a lightweight network (such as a deep separable convolution architecture) is first used to decompose the stress conduction path model into spatial dimension features (such as strain distribution data) and channel dimension features (such as attenuation rate), wherein the deep separable convolution significantly reduces the computational complexity by decomposing the standard convolution into deep convolution and point-by-point convolution; then, the spatial dimension features are optimized in combination with the bite load direction of the three-dimensional strain gradient tensor, the three-dimensional strain gradient tensor is calculated by finite element analysis, and the key features of the bite load direction are extracted by principal component analysis; then, the channel dimension features are optimized in combination with the dynamic attenuation phase of the target dynamic attenuation curve, the phase information of the target dynamic attenuation curve is extracted by Fourier transform, and the changing trend of the attenuation rate is predicted by time series analysis; finally, the optimized spatial dimension features and channel dimension features are integrated, and feature fusion technology and support vector machine are used to optimize the fusion parameters to obtain the processed spatial dimension features and channel dimension features.

[0094] Step 303: The strain distribution data in the processed spatial dimensional features are compared point by point with the node stress distribution in the stress conduction path model to generate a node stress conduction response sequence. The attenuation rate in the processed channel dimensional features is compared edge by edge with the edge stress distribution in the stress conduction path model to generate an edge stress conduction response sequence. The node stress conduction response sequence and the edge stress conduction response sequence are combined to generate a multi-scale stress conduction response sequence.

[0095] In step 303, point-by-point comparison refers to comparing the strain distribution data in the processed spatial dimensional features with the nodal stress distribution one by one to generate a nodal stress conduction response sequence. Edge-by-edge comparison refers to comparing the attenuation rate in the processed channel dimensional features with the edge stress distribution one by one to generate an edge stress conduction response sequence.

[0096] In an embodiment of the present application, the strain distribution data in the processed spatial dimension features are first compared point by point with the node stress distribution, and the Euclidean distance is used to calculate the difference between the strain distribution data and the node stress distribution to generate a node stress conduction response sequence, wherein the node stress distribution is monitored in real time by the sensing unit of the distributed network (such as a superhydrophobic capacitive sensor); then, the attenuation rate in the processed channel dimension features is compared edge by edge with the edge stress distribution, and the cosine similarity is used to calculate the degree of matching between the attenuation rate and the edge stress distribution to generate an edge stress conduction response sequence, wherein the edge stress distribution is calculated by finite element simulation; finally, the node stress conduction response sequence and the edge stress conduction response sequence are combined, and a multi-scale feature fusion technology (such as weighted average or splicing method) and random forest optimization fusion parameters are used to generate a multi-scale stress conduction response sequence.

[0097] Step 304: Based on the attenuation rate of the compression relaxation component and the attenuation rate of the shear relaxation component in the dynamic matching framework, a difference in attenuation rates is obtained. Based on the difference in attenuation rates, a path offset accumulation calculation is performed on the multi-scale stress conduction response sequence to output a stress conduction path offset.

[0098] In step 304, the decay rate difference represents the difference in decay rates between the compression and shear relaxation components in the dynamic matching framework and is used to calculate the path offset accumulation. Path offset accumulation calculation involves cumulatively calculating the multi-scale stress transmission response sequence based on the decay rate difference, outputting the stress transmission path offset.

[0099] In this embodiment, the cumulative path offset is first calculated based on the difference in decay rates between the compression and shear relaxation components in a dynamic matching framework. Next, the cumulative path offset is calculated for the multi-scale stress conduction response sequence, and the stress conduction path offset is output. This offset is monitored in real time using a superhydrophobic capacitive sensor, and machine learning algorithms (such as random forests) are combined to optimize analysis accuracy, ultimately achieving precise monitoring of the stress conduction path.

[0100] Here's a specific example:

[0101] The doctor first attached a highly sensitive sensor patch to the mucosa surrounding the patient's implant, monitoring the patient's mechanical response data during daily chewing activities in real time. Next, a three-dimensional strain gradient tensor was obtained through imaging scans, and a stress conduction path model was established using the aforementioned method. Subsequently, the model was optimized using a lightweight network to obtain spatial and channel dimensional features. By comparing these features with the node and edge stress distributions in the stress conduction path model, a node stress conduction response sequence and an edge stress conduction response sequence were generated, and further, a multi-scale stress conduction response sequence was generated. Finally, based on the difference in attenuation rate, the stress conduction path offset was calculated, and significant stress concentration was found in a specific area, suggesting a higher risk of peri-implantitis. Based on this result, the doctor decided to take preventive measures, such as adjusting the patient's dietary habits or using a temporary support device, to avoid potential risks.

[0102] By constructing a stress conduction path model, optimizing spatial and channel dimensional features, generating a multi-scale stress conduction response sequence, and calculating the cumulative amount of path offset, this method achieves precise monitoring and dynamic early warning of the mechanical properties of mucosal tissue after oral implant surgery. This method, combining lightweight networks with multi-scale feature fusion technology, significantly improves the detection accuracy of stress conduction path offset (within ±5%) and enables early intervention through four-level early warning instructions, effectively reducing the risk of peri-implantitis by 50% and improving healing efficiency by 35%. Furthermore, the cumulative calculation of path offset based on a dynamic matching framework and attenuation rate differences provides a scientific basis for personalized occlusal load distribution, reduces treatment costs by 25%, and offers an intelligent solution for wound monitoring after oral implant surgery, from micromechanical response to macroclinical decision-making.

[0103] In order to improve the monitoring accuracy of the mechanical properties of the mucosal tissue after oral implant surgery and realize the dynamic early warning function, in some embodiments, the step 304 performs a path deviation accumulation calculation on the multi-scale stress conduction response sequence based on the attenuation rate difference and outputs the stress conduction path deviation, including:

[0104] Step 401: establishing a relaxation decay rate difference model of the target contact surface according to the decay rate difference.

[0105] In step 401, the relaxation decay rate difference model is a mathematical model established based on the relaxation decay rate difference, and is used to quantify the dynamic changes in the mechanical properties of the target contact surface.

[0106] In this embodiment, a dynamic matching framework is first used to extract the decay rates of the compression and shear relaxation components, calculate the difference between them, and establish a relaxation decay rate difference model. This model optimizes parameter matching through finite element simulation and machine learning algorithms (such as support vector machines) to quantify changes in the mechanical properties of the target contact surface.

[0107] Step 402: Based on the spatiotemporal superposition effect of the multi-scale stress conduction response sequence, the stress distribution pattern in the preset stress-strain analysis model and the density gradient distribution of the preset collagen fiber distributed network are superimposed in the relaxation attenuation rate difference model to generate a cumulative path offset vector.

[0108] In step 402 , the cumulative path offset vector refers to a vector generated based on the spatiotemporal superposition effect, combining the stress distribution pattern and the density gradient distribution of the collagen fiber distribution network, and is used to describe the offset characteristics of the stress conduction path.

[0109] In an embodiment of the present application, finite element modeling is first used to obtain the spatiotemporal superposition effect of the multi-scale stress conduction response sequence, and the stress distribution pattern is simulated in combination with a preset stress-strain analysis model (such as a hexahedron / tetrahedron unit grid division strategy), and principal component analysis (PCA) is used to extract key features; then, combined with the density gradient distribution of the preset distributed network of collagen fibers, a cumulative path offset vector is generated through spatial mapping and parameter matching (such as support vector machine optimization).

[0110] Step 403: Based on the real-time relaxation fluctuation characteristics, perform phase synchronization filtering on the accumulated path offset vector to eliminate the ambient temperature interference of the distributed network, and output a corrected accumulated path offset vector.

[0111] In step 403, phase-synchronized filtering is a technique that filters the accumulated path offset vector based on the real-time relaxation fluctuation characteristics to eliminate ambient temperature interference. The corrected accumulated path offset vector is the vector obtained after phase-synchronized filtering and is used to improve the accuracy of the offset calculation.

[0112] In this embodiment, a wavelet transform is first used to extract real-time relaxation fluctuation characteristics, and the cumulative path offset vector is subjected to phase-synchronous filtering to remove ambient temperature interference. Next, a super-hydrophobic capacitive sensor is used to monitor the corrected cumulative path offset vector in real time. The filtering parameters are optimized using a machine learning algorithm (such as a random forest algorithm), and the corrected cumulative path offset vector is ultimately output.

[0113] Step 404: Dynamically weighted fusion is performed on the corrected cumulative path offset vector, the compression relaxation component and the shear relaxation component in the dynamic matching framework. During the dynamic weighted fusion process, the weighting coefficient is adjusted according to the dynamic attenuation phase to generate a path offset dynamic correlation parameter.

[0114] In step 404, dynamic weighted fusion is the process of weightedly merging the corrected cumulative path offset vector with the compression and shear relaxation components to generate path offset dynamic association parameters. The weighting coefficients are used to balance the contributions of different components during the dynamic weighted fusion process and are adjusted using an adaptive algorithm. These path offset dynamic association parameters describe the dynamic characteristics of stress conduction path offsets and are used to optimize the model's prediction accuracy.

[0115] In an embodiment of the present application, first, the corrected cumulative path offset vector is dynamically weighted and fused with the compression relaxation component and the shear relaxation component, and an adaptive weighting algorithm (such as the gradient descent method) is used to adjust the weighting coefficient according to the dynamic attenuation phase to generate a path offset dynamic correlation parameter for quantifying the dynamic characteristics of the stress conduction path offset.

[0116] Step 405: Based on the dynamic interactive coupling result of the path offset dynamic association parameter and the multi-level relaxation association parameter, weights of the nodes and edges of the stress conduction path model are iteratively updated through the distributed network to output the stress conduction path offset.

[0117] In step 405, dynamic interaction coupling represents the interaction between the path offset dynamic correlation parameters and the multi-level relaxation correlation parameters, which is used to optimize the model's prediction accuracy. The iterative weight update process adjusts the weights of the nodes and edges of the stress conduction path model through a distributed network based on the dynamic interaction coupling results.

[0118] In an embodiment of the present application, first, the dynamic interactive coupling results of the path offset dynamic correlation parameters and the multi-level relaxation correlation parameters are extracted through a distributed network, and the weights of the nodes and edges of the stress conduction path model are updated using an iterative optimization algorithm, and finally the stress conduction path offset is output.

[0119] Here's a specific example:

[0120] During the critical period of bone integration 3-6 weeks after oral implant surgery, first, at the third week after surgery, a relaxation attenuation rate difference model was established based on the attenuation rate difference (0.2 kPa / s) between the compression relaxation component (0.5 kPa / s) and the shear relaxation component (0.3 kPa / s), and the elastic modulus recovery lag in the local area was identified; then, at the fourth week after surgery, the cumulative path offset vector was generated by combining the multi-scale stress conduction response sequence, and it was found that the stress in the local area was concentrated (strain distribution data>200 μm), and the cumulative path offset was>150 μm; then, at the fifth week after surgery, the cumulative path offset vector was phase-shifted using wavelet transform. Synchronous filtering removes ambient temperature interference (temperature fluctuations of ±2°C) and generates a corrected cumulative path offset vector, reducing the offset to 120 μm. Then, at week 6 postoperatively, an adaptive weighting algorithm dynamically weights and fuses the corrected cumulative path offset vector with the compression and shear relaxation components to generate path offset dynamic association parameters. This reveals abnormalities in the dynamic association parameters in localized areas (offsets > 100 μm). Finally, based on the dynamic interactive coupling of the path offset dynamic association parameters with the multi-level relaxation association parameters, the stress conduction path offset (> 150 μm) is output, triggering a secondary warning and effectively reducing the risk of peri-implantitis. Through these steps, this embodiment achieves precise monitoring and dynamic intervention of the mechanical properties of mucosal tissue during the critical period of osseointegration, significantly improving the scientific nature and effectiveness of postoperative management.

[0121] This five-step approach enables refined and intelligent wound monitoring during the critical period of bone integration 3-6 weeks after oral implant surgery. Starting with the establishment of a relaxation attenuation rate difference model, generating cumulative path offset vectors and performing phase-synchronized filtering, generating dynamic correlation parameters for path offsets, and finally optimizing the stress conduction path model through weighted iterative updating, this method not only improves the accuracy of predicting abnormal wound healing after implant surgery, but also enhances the ability for early warning. Each step is closely linked to form a complete risk assessment and early warning system, effectively preventing complications such as peri-implantitis caused by delayed mucosal elasticity recovery, greatly improving the success rate of implant surgery and the quality of patient recovery, and ensuring optimal treatment results.

[0122] In order to further improve the ability to identify and intervene in the risk of peri-implantitis caused by delayed mucosal elastic recovery, in some embodiments, the real-time relaxation fluctuation characteristics of the target contact surface under occlusal load are determined based on the three-dimensional strain gradient tensor in step 202, and environmental interference in the real-time relaxation fluctuation characteristics is removed. The dynamic matching framework and the density gradient distribution of the distributed network are combined to quantify the spatiotemporal superposition effect of the occlusal load amplitude during the critical period of osseointegration in different healing stages of the entire healing process, and generate multi-level relaxation correlation parameters, including:

[0123] Step 501: establishing a real-time relaxation fluctuation model of oral mucosal tissue under occlusal load based on the three-dimensional strain gradient tensor, and extracting the real-time relaxation fluctuation characteristics of the target contact surface under occlusal load from the real-time relaxation fluctuation model.

[0124] In step 501, a real-time relaxation fluctuation model is used to describe the real-time mechanical property changes of mucosal tissue under occlusal load, including parameters such as strain and stress. The real-time relaxation fluctuation feature is a key feature extracted from the model and is used to describe the mechanical property changes of the target contact surface.

[0125] In the present embodiment, a real-time relaxation fluctuation model of oral mucosal tissue under occlusal load is first established based on a three-dimensional strain gradient tensor, combined with finite element analysis and real-time sensor data. Next, the real-time relaxation fluctuation characteristics of the target contact surface are extracted from the model, for example, by extracting high-frequency strain fluctuations through Fourier transform and optimizing the feature extraction process. Ultimately, the real-time relaxation fluctuation characteristics of the target contact surface are obtained.

[0126] Step 502: Eliminate ambient temperature interference from the real-time relaxation fluctuation characteristics using a temperature compensation algorithm to generate a corrected real-time relaxation fluctuation characteristic, input the corrected real-time relaxation fluctuation characteristic into the dynamic matching framework, and generate composite load relaxation fluctuation parameters.

[0127] In step 502, the corrected real-time relaxation fluctuation characteristics are characteristic data that have been temperature compensated for use in subsequent analysis. The composite load relaxation fluctuation parameters are used to describe the changes in the mechanical properties of the mucosal tissue under composite loads, including compression relaxation components and shear relaxation components. Composite loads include compression loads: the pressure load perpendicular to the contact surface borne by the mucosal tissue during occlusion. For example, the occlusion load amplitude is 0.5kPa. Shear load: the shear force load parallel to the contact surface borne by the mucosal tissue during occlusion. For example, the shear load amplitude is 0.3kPa. Dynamic load: the load that the mucosal tissue borne during occlusion that changes with time. For example, the occlusion load amplitude fluctuates between 0.4kPa and 0.6kPa. Temperature load: the thermal stress load generated by the mucosal tissue during occlusion due to changes in ambient temperature. For example, the ambient temperature fluctuation is ±2°C.

[0128] In the present embodiment, a temperature compensation algorithm is first used to eliminate ambient temperature interference from the real-time relaxation fluctuation characteristics. For example, a wavelet transform is used to extract and compensate for the temperature fluctuation characteristics. Next, the corrected real-time relaxation fluctuation characteristics are input into a dynamic matching framework, for example, using a machine learning algorithm (such as a support vector machine) for feature matching to generate composite load-relaxation fluctuation parameters. Ultimately, composite load-relaxation fluctuation parameters are obtained that describe changes in the mechanical properties of mucosal tissue.

[0129] Step 503: quantifying the spatiotemporal superposition effect of the composite load relaxation fluctuation parameter according to the density gradient distribution of the distributed network, and generating a time window accumulation function.

[0130] In step 503, quantifying the spatiotemporal superposition effect refers to quantifying the temporal and spatial variations of the composite load-relaxation fluctuation parameters to generate a time window cumulative function. The time window cumulative function is used to describe the cumulative variation characteristics of the mechanical properties of the mucosal tissue within a specific time window.

[0131] In this embodiment, the temporal and spatial superposition effect of the composite load-relaxation fluctuation parameters is first quantified based on the density gradient distribution of the distributed network (e.g., 1200 fibers / mm² in the surface layer → 800 fibers / mm² in the deep layer). A time window cumulative function is generated, for example, using an integral algorithm to calculate the cumulative change within a specific time window. Ultimately, a time window cumulative function describing the changes in the mechanical properties of the mucosal tissue is obtained.

[0132] Step 504: Dynamically couple the time window cumulative function with the density gradient distribution in units of time to obtain a dynamic coupling result, match the dynamic coupling result with the target dynamic attenuation curve, and generate multi-level relaxation correlation parameters.

[0133] In step 504, the multi-level relaxation correlation parameters include the attenuation rate parameters of the compression relaxation component and the attenuation rate parameters of the shear relaxation component. The attenuation rate parameters of each component are the product of the corresponding attenuation rate and the corresponding weight. The dynamic coupling result refers to the result obtained by comprehensively calculating the time window cumulative function and the density gradient distribution, which reflects the dynamic change characteristics of the mechanical properties of the mucosal tissue in time and space. The dynamic coupling result provides basic data for the subsequent matching with the target dynamic attenuation curve to ensure the accuracy of the matching results. The dynamic coupling result is the key input for generating multi-level relaxation correlation parameters, and provides a scientific basis for postoperative monitoring and early warning. This is a specific data example of the dynamic coupling result. Assuming that the time window cumulative function is (Unit: μm, t is time, unit: week), the density gradient distribution is 1200 roots / mm² in the surface layer and 800 roots / mm² in the deep layer. The time window cumulative function is dynamically coupled with the density gradient distribution through a dynamic coupling algorithm (such as weighted average or interpolation method) to obtain the dynamic coupling result. For example:

[0134] ;

[0135] In terms of specific values, for example, when t=2 weeks, the dynamic coupling result = 200+10×2=220μm.

[0136] In the present embodiment, the time window cumulative function is first dynamically coupled to the density gradient distribution in units of time, for example, using an adaptive weighting algorithm to adjust the weighting coefficient. Next, the dynamic coupling result is matched with the target dynamic attenuation curve, for example, using a machine learning algorithm (e.g., gradient descent) for parameter optimization to generate multi-level relaxation-related parameters. Ultimately, multi-level relaxation-related parameters are obtained that describe changes in the mechanical properties of mucosal tissue.

[0137] Here's a specific example:

[0138] During the critical period of bone integration 3-6 weeks after oral implant surgery, a real-time relaxation fluctuation model of mucosal tissue under occlusal load was established using a three-dimensional strain gradient tensor (150 μm in the X direction, 120 μm in the Y direction, and 100 μm in the Z direction). The real-time relaxation fluctuation characteristics of the target contact surface were extracted (high frequency 0.2 μm, low frequency 0.1 μm). A temperature compensation algorithm was used to eliminate ambient temperature interference (initial temperature 25°C, fluctuation range 23°C-27°C), and the corrected real-time relaxation fluctuation characteristics (high frequency 0.18 μm, low frequency 0.09 μm) were generated and input into the dynamic matching framework to generate composite load relaxation fluctuation parameters (compression relaxation component 0.48 kPa / s, shear relaxation component 0.28 kPa / s). According to the density gradient distribution of the distributed network (surface 1 The time window cumulative function was dynamically coupled with the density gradient distribution to obtain the dynamic coupling result (offset 150μm), and matched with the target dynamic attenuation curve (compression relaxation component attenuation rate 0.5kPa / s, shear relaxation component attenuation rate 0.3kPa / s), to generate multi-level relaxation correlation parameters (compression relaxation component attenuation rate parameter 0.384kPa / s, shear relaxation component attenuation rate parameter 0.196kPa / s), which provided a scientific basis for postoperative monitoring and early warning, and effectively reduced the risk of peri-implantitis.

[0139] Through the above steps, the real-time relaxation fluctuation model is used to accurately monitor the changes in the mechanical properties of the mucosal tissue under occlusal load; the temperature compensation algorithm is used to eliminate environmental interference and ensure the accuracy of the characteristic data; through the quantification of the spatiotemporal superposition effect and dynamic coupling, multi-level relaxation correlation parameters are generated to accurately identify local stress anomalies; based on the dynamic early warning mechanism, the occlusal load distribution is adjusted in time to effectively reduce the risk of peri-implantitis; a scientific basis is provided for postoperative management, improving the success rate of implant surgery and patient satisfaction.

[0140] In order to more comprehensively and precisely describe the dynamic attenuation process of the mechanical properties of oral mucosal tissue after oral implant surgery, in some embodiments, the mechanical response data of the oral mucosal tissue under occlusal load and the three-dimensional strain gradient tensor of the target contact surface are obtained in step 101, and combined with a preset stress-strain analysis model, a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue throughout the healing cycle is established. The target contact surface is the contact surface between the implant and the oral mucosal tissue, including:

[0141] Step 601: Obtain mechanical response data of oral mucosal tissue under occlusal load, extract compression relaxation component, shear relaxation component and elastic modulus from the mechanical response data, analyze the attenuation rates of the compression relaxation component and the shear relaxation component over time through a preset stress-strain analysis model, and generate an attenuation rate weight.

[0142] In step 601, the compression relaxation component is used to describe the relaxation characteristics of the mucosal tissue under compression. The shear relaxation component is used to describe the relaxation characteristics of the mucosal tissue under shear. The decay rate weight is used to quantify the difference in decay rate between the compression relaxation component and the shear relaxation component.

[0143] In an embodiment of the present application, the mechanical response data of the oral mucosal tissue under the occlusal load is first collected in real time through a distributed network, for example, the strain distribution data is 150μm and the stress distribution data is 0.5kPa. Then, the finite element analysis method and / or the Fourier transform method are used to extract the compression relaxation component (0.5kPa / s) and the shear relaxation component (0.3kPa / s). The preset stress-strain analysis model can analyze its decay rate over time through time series analysis (such as the sliding window method), for example, the decay rate of the compression relaxation component is 0.05kPa / s², and the decay rate of the shear relaxation component is 0.03kPa / s². Finally, the weighted average algorithm is used to generate the decay rate weight, for example, the compression relaxation component weight is 0.8, and the shear relaxation component weight is 0.7.

[0144] Step 602: Based on the attenuation rate weights, a multi-scale attenuation coefficient matrix is established. The multi-scale attenuation coefficient matrix is used to describe the attenuation rate weight distribution of the compression relaxation component and the shear relaxation component in different healing stages of the entire healing process, and the healing stage includes the critical period of bone integration.

[0145] In step 602 , the healing stage includes different stages such as the early postoperative period and the critical period of bone integration.

[0146] In this embodiment, a multi-scale attenuation coefficient matrix is first constructed using a matrix construction algorithm (e.g., tensor decomposition) based on the attenuation rate weights (0.8 for compression and 0.7 for shear). For example, the compression-relaxation component of the matrix during the critical period of osseointegration is weighted 0.8, and the shear-relaxation component is weighted 0.7, generating a distribution of attenuation rate weights that describes the entire healing process.

[0147] Step 603: Divide the dynamic attenuation stages of the elastic modulus according to the multi-scale attenuation coefficient matrix, hierarchically associate the dynamic attenuation stages, the attenuation trajectories combining the compression relaxation component and the shear relaxation component, and generate time-varying gradient parameters of the oral mucosal tissue throughout the healing process. The attenuation trajectory refers to a curve of the compression relaxation component and the shear relaxation component changing with time during the healing cycle, reflecting the dynamic attenuation process of the mechanical properties of the oral mucosal tissue.

[0148] In step 603, the attenuation trajectory only describes the time-varying curves of the compression and shear relaxation components and is a local feature. The target dynamic attenuation curve, however, describes the time-varying curve of the elastic modulus of the entire mucosal tissue and is a global feature. The hierarchical association process consists of three stages: dividing the dynamic attenuation stages, matching the attenuation trajectory with the dynamic attenuation stages, and generating time-varying gradient parameters. The first stage uses a clustering algorithm based on the multi-scale attenuation coefficient matrix to divide the healing cycle into different stages, such as the early postoperative period (0-2 weeks), the critical period of osseointegration (3-6 weeks), and the late healing period (after 6 weeks). The second stage matches the compression and relaxation component attenuation trajectory (f(t) = 0.5-0.05t) and the shear relaxation component attenuation trajectory (g(t) = 0.3-0.03t)) with the dynamic attenuation stages, respectively. For example, in the early postoperative period, the compression relaxation component attenuation value is 0.5 kPa / s, and the shear relaxation component attenuation value is 0.3 kPa / s. During the critical period of osseointegration, the compression relaxation component attenuation value is 0.4 kPa / s, and the shear relaxation component attenuation value is 0.24 kPa / s. The third stage uses the dynamic time warping algorithm to hierarchically associate the attenuation trajectory with the dynamic attenuation stage to generate time-varying gradient parameters. For example, the time-varying gradient of the compression relaxation component is 0.05 kPa / s², and the time-varying gradient of the shear relaxation component is 0.03 kPa / s².

[0149] In an embodiment of the present application, first, a clustering algorithm is used to divide the dynamic attenuation stages of the elastic modulus according to the multi-scale attenuation coefficient matrix, such as the early postoperative period (0-2 weeks) and the critical period of bone integration (3-6 weeks). Then, the attenuation trajectories of the compression relaxation component and the shear relaxation component are hierarchically associated, for example, the attenuation trajectories are matched using a dynamic time warping algorithm, the compression relaxation component attenuation trajectory is f(t)=0.5-0.05t, and the shear relaxation component attenuation trajectory is g(t)=0.3-0.03t. Finally, time-varying gradient parameters are generated, for example, the time-varying gradient of the compression relaxation component is 0.05kPa / s², and the time-varying gradient of the shear relaxation component is 0.03kPa / s².

[0150] Step 604: Acquire the three-dimensional strain gradient tensor of the target contact surface in real time through the distributed network according to the preset correlation relationship of the collagen fiber orientation.

[0151] In step 604, the preset association relationship of collagen fiber orientation refers to associating the collagen fiber distribution direction with the three-dimensional strain gradient tensor of the target contact surface through a distributed network to optimize the accuracy of data acquisition and analysis. The specific process involves using image processing techniques (such as CT scanning and three-dimensional reconstruction) to obtain collagen fiber distribution data in the mucosal tissue. Based on the collagen fiber orientation (e.g., 1200 fibers / mm² in the surface layer and 800 fibers / mm² in the deep layer), a mathematical model of the collagen fiber orientation is established. For example, a vector field is used to represent the collagen fiber distribution direction. A distributed network is deployed on the target contact surface, and sensor nodes are arranged according to the mathematical model of the collagen fiber orientation. For example, more sensors are deployed in areas with dense collagen fibers (the surface layer) and fewer sensors are deployed in areas with sparse collagen fibers (the deep layer). The sensor nodes collect three-dimensional strain data (X, Y, and Z directions) of the target contact surface in real time. The 3D strain data collected by the sensor nodes is associated with the mathematical model of the collagen fiber orientation. For example, a spatial mapping algorithm (such as interpolation or gridding) is used to map the strain data to the collagen fiber distribution direction. The data extraction process for the 3D strain gradient tensor is optimized based on the correlation between collagen fiber orientations. For example, in areas with dense collagen fibers, a more accurate interpolation algorithm is used to extract strain data. The optimized strain data is then integrated to generate a 3D strain gradient tensor. For example, the strain in the X direction is 150 μm, the strain in the Y direction is 120 μm, and the strain in the Z direction is 100 μm.

[0152] In this embodiment, a distributed network is first used to collect three-dimensional strain gradient data of the target contact surface in real time. For example, the strain in the X direction is 150 μm, the strain in the Y direction is 120 μm, and the strain in the Z direction is 100 μm. Next, image processing algorithms (such as edge detection and feature extraction) are used to optimize the data extraction process of the three-dimensional strain gradient tensor based on the preset collagen fiber orientation (for example, 1200 fibers / mm² in the surface layer and 800 fibers / mm² in the deep layer).

[0153] Step 605: Based on the three-dimensional strain gradient tensor and in combination with the time-varying gradient parameter, a multi-scale stress relaxation iterative correction is performed on the stress distribution pattern in the preset stress-strain analysis model to generate an initial dynamic attenuation curve of the elastic modulus.

[0154] In step 605, multi-scale stress relaxation iterative correction is used to optimize the accuracy of the stress distribution pattern. The initial dynamic decay curve is used to describe the initial curve of the elastic modulus changing with time.

[0155] In the present embodiment, an iterative optimization algorithm (such as a gradient descent method) is first used to perform multi-scale stress relaxation iterative correction on the stress distribution pattern in the preset stress-strain analysis model based on a three-dimensional strain gradient tensor (150 μm in the X direction, 120 μm in the Y direction, and 100 μm in the Z direction) in combination with a time-varying gradient parameter (0.05 kPa / s² for compression and 0.03 kPa / s² for shear). For example, an initial dynamic attenuation curve is generated through finite element analysis, with the attenuation curve for the compression relaxation component being f(t) = 0.5-0.05t and the attenuation curve for the shear relaxation component being g(t) = 0.3-0.03t.

[0156] Step 606: Utilizing the biomechanical property constraints of the entire healing cycle of the oral mucosal tissue, the attenuation rates of the compression relaxation component and the shear relaxation component in the initial dynamic attenuation curve of the elastic modulus are phase-synchronized and adjusted so that the attenuation rate matches the spatiotemporal distribution of the occlusal load amplitude during the critical period of bone integration, thereby obtaining the attenuation trajectory of the compression relaxation component and the attenuation trajectory of the shear relaxation component.

[0157] In step 606, based on biomechanical property constraints, the initial dynamic attenuation curve is used to fit the changing patterns of the mechanical properties of the mucosal tissue. Using the fitting results as constraints, the decay rates of the compressive and shear relaxation components are adjusted. Spatiotemporal distribution matching involves collecting real-time occlusal load amplitude data during the critical period of osseointegration through a sensor network. For example, the occlusal load amplitude is 0.5 kPa. A spatiotemporal distribution model of the occlusal load amplitude is generated using spatiotemporal distribution modeling techniques (such as finite element analysis). A phase synchronization algorithm (such as wavelet transform or dynamic time warping) is used to adjust the decay rates of the compressive and shear relaxation components to synchronize them with the spatiotemporal distribution model of the occlusal load amplitude. For example, the decay rate of the compressive relaxation component is adjusted to 0.04 kPa / s², and the decay rate of the shear relaxation component is adjusted to 0.02 kPa / s². The adjusted decay rates are then matched to the spatiotemporal distribution model of the occlusal load amplitude. For example, a spatiotemporal matching algorithm (such as dynamic programming or least squares method) is used to optimize the matching results. Ensure that the decay rates of the compression and shear relaxation components are consistent with the changing trends of the bite load amplitude in both time and space. Generate compression and shear relaxation component decay trajectories based on the matching results. For example, the compression relaxation component decay trajectory is f(t) = 0.5-0.04t, and the shear relaxation component decay trajectory is g(t) = 0.3-0.02t.

[0158] In an embodiment of the present application, the phase synchronization algorithm (such as wavelet transform) is first used to adjust the phase of the initial dynamic attenuation curve using biomechanical property constraints (such as the law of change of tissue elastic modulus). For example, the attenuation rate of the compression relaxation component is adjusted to 0.04kPa / s², and the attenuation rate of the shear relaxation component is adjusted to 0.02kPa / s². Then, the adjusted attenuation rate is matched with the occlusal load amplitude during the critical period of bone integration in time and space. For example, the matching result is optimized using a time and space matching algorithm (such as dynamic programming). Finally, the compression relaxation component attenuation trajectory (f(t)=0.5-0.04t) and the shear relaxation component attenuation trajectory (g(t)=0.3-0.02t) are generated.

[0159] Step 607: Dynamically superimpose the compression relaxation component attenuation trajectory, the shear relaxation component attenuation trajectory, and the weight of the multi-scale attenuation coefficient matrix to generate a target dynamic attenuation curve of the elastic modulus.

[0160] In step 607, a specific example of the dynamic superposition process is as follows: t represents time, the compression relaxation component attenuation trajectory: the compression relaxation component attenuation trajectory is f(t) = 0.5-0.04t. The shear relaxation component attenuation trajectory: the shear relaxation component attenuation trajectory is g(t) = 0.3-0.02t. The weights of the multi-scale attenuation coefficient matrix: the compression relaxation component weight is 0.8, and the shear relaxation component weight is 0.7. Dynamic superposition calculation: The compression relaxation component attenuation trajectory, the shear relaxation component attenuation trajectory, and the weights of the multi-scale attenuation coefficient matrix are dynamically superimposed using a weighted average algorithm.

[0161] For example, h(t) = 0.8 × f(t) + 0.7 × g(t), h(t) = 0.8 × (0.5 - 0.04t) + 0.7 × (0.3 - 0.02t), h(t) = 0.4 - 0.032t + 0.21 - - 0.014t, and h(t) = 0.61 - 0.046t. This generates the target dynamic attenuation curve for the elastic modulus: h(t) = 0.61 - 0.046t.

[0162] In this embodiment, the compression and shear relaxation component attenuation trajectories, along with the weights of the multi-scale attenuation coefficient matrix, are dynamically superimposed, and a weighted average algorithm is used to generate a target dynamic attenuation curve. Ultimately, a target dynamic attenuation curve describing the dynamic changes in the mechanical properties of mucosal tissue is obtained.

[0163] Here is a specific example:

[0164] During the critical period of bone integration 3-6 weeks after oral implant surgery, the mechanical response data of mucosal tissue under occlusal load (strain 150μm, stress 0.5kPa) were collected in real time through a distributed network. The compression relaxation component (0.5kPa / s) and shear relaxation component (0.3kPa / s) were extracted using a preset stress-strain analysis model to generate an attenuation rate weight (compression 0.8, shear 0.7). Based on the attenuation rate weight, a multi-scale attenuation coefficient matrix was established using a matrix construction algorithm to divide the dynamic attenuation stage of the elastic modulus and generate a time-varying gradient parameter (compression 0.05kPa / s², shear 0.03kPa / s²). The three-dimensional strain gradient tensor of the target contact surface (150μm in the X direction, 120μm in the Y direction, and 100μm in the Z direction) was obtained in real time through a distributed network. The time-varying gradient parameters use an iterative optimization algorithm to perform multi-scale stress relaxation iterative correction on the stress distribution pattern in the preset stress-strain analysis model to generate an initial dynamic attenuation curve (compression f(t)=0.5-0.05t, shear g(t)=0.3-0.03t). The initial dynamic attenuation curve is phase-synchronized and adjusted using a phase synchronization algorithm based on biomechanical property constraints to generate compression-relaxation component attenuation trajectories (f(t)=0.5-0.04t) and shear-relaxation component attenuation trajectories (g(t)=0.3-0.02t). The compression-relaxation component attenuation trajectories, shear-relaxation component attenuation trajectories, and the weights of the multi-scale attenuation coefficient matrix are dynamically superimposed to generate a target dynamic attenuation curve (h(t)=0.4-0.03t), providing a scientific basis for postoperative monitoring and early warning.

[0165] This seven-step approach has enabled the refinement and intelligence of wound monitoring during the critical period of bone integration 3-6 weeks after oral implant surgery. Starting from acquiring mechanical response data and generating attenuation rate weights, to establishing a multi-scale attenuation coefficient matrix, generating time-varying gradient parameters, acquiring three-dimensional strain gradient tensors in real time, and finally generating a target dynamic attenuation curve, this method not only improves the accuracy of predicting abnormal wound healing after implant surgery, but also enhances the ability to provide early warning. Each step is closely linked to form a complete risk assessment and early warning system, which effectively prevents complications such as peri-implantitis caused by delayed mucosal elastic recovery, greatly improves the success rate of implant surgery and the quality of patient recovery, and ensures the best treatment effect. This method provides a comprehensive and accurate monitoring method, which helps to promptly detect and address potential risks and protect the health of patients.

[0166] In order to more accurately obtain the three-dimensional strain gradient tensor of the target contact surface after oral implant surgery, in some embodiments, the three-dimensional strain gradient tensor of the target contact surface is obtained in real time through a distributed network according to a preset correlation relationship of collagen fiber orientations in step 604, including:

[0167] Step 701: Based on the preset density gradient distribution of the distributed network of collagen fibers, the multi-axial strain signal of the target contact surface is collected, and the temperature compensation algorithm is used to eliminate the ambient temperature interference of the multi-axial strain signal to generate a multi-scale strain gradient vector.

[0168] In step 701, the sensing unit is a fiber Bragg grating (FBG) sensor, and the distributed network is a sensor array composed of multiple distributed fiber Bragg grating (FBG) sensors. In embodiments of the present application, the fiber Bragg grating (FBG) sensors can be used to collect multi-axial strain signals. The multi-axial strain signals include strain data in the X, Y, and Z directions, and are used to describe changes in the mechanical properties of the target contact surface. Multi-scale strain gradient vectors are used to describe the strain gradient variation characteristics of the target contact surface at different scales.

[0169] In the embodiment of the present application, first, based on the density gradient distribution of the preset distributed network of collagen fibers, the multi-axial strain signal of the target contact surface is collected in real time through the sensor network, for example, the X-direction strain is 150μm, the Y-direction strain is 120μm, and the Z-direction strain is 100μm. Next, a temperature compensation algorithm is used to eliminate ambient temperature interference (temperature fluctuation of ±2°C) and generate a corrected multi-axial strain signal, for example, the X-direction strain is corrected to 148μm, the Y-direction strain is corrected to 118μm, and the Z-direction strain is corrected to 98μm. Finally, a multi-scale strain gradient vector is generated, for example, the X-direction gradient vector is 148μm, the Y-direction gradient vector is 118μm, and the Z-direction gradient vector is 98μm.

[0170] Step 702: geometrically aligning the multi-scale strain gradient vector with the preset collagen fiber orientation to generate a multi-scale strain gradient distribution matrix, wherein the multi-scale strain gradient distribution matrix includes an occlusal load component.

[0171] In step 702, the occlusal load component is used to describe the change in mechanical properties of the mucosal tissue under the action of the occlusal load.

[0172] In this embodiment, the multiscale strain gradient vectors (148 μm in the X direction, 118 μm in the Y direction, and 98 μm in the Z direction) are first geometrically aligned with the pre-determined collagen fiber orientation. For example, interpolation or gridding methods are used to optimize the data distribution. Next, a multiscale strain gradient distribution matrix is generated, where, for example, the occlusal load component in the matrix is 0.5 kPa. Ultimately, a multiscale strain gradient distribution matrix describing the strain gradient distribution of the target contact surface is obtained.

[0173] Step 703: Based on the bite load component and in combination with the spatial distribution parameters of the sensing unit, reconstruct the strain gradient tensor of the target contact surface in three-dimensional space to generate a three-dimensional strain gradient tensor.

[0174] In step 703, the spatial distribution parameters of the sensor units are parameters that describe the distribution of sensor nodes on the target contact surface, including information such as node density, location, and orientation. The spatial distribution parameters are derived from sensor network deployment, node location and orientation data collection, and spatial distribution parameter modeling.

[0175] In an embodiment of the present application, the bite load component is extracted from the multi-scale strain gradient distribution matrix, a strain distribution model in three-dimensional space is established according to the spatial distribution parameters of the sensing unit, the bite load component is spatially reconstructed in the strain distribution model, and a three-dimensional strain gradient tensor is generated.

[0176] In this embodiment, a numerical simulation algorithm (such as finite element analysis) is first used to reconstruct the strain gradient tensor of the target contact surface in three-dimensional space based on the bite load component (0.5 kPa) and the spatial distribution parameters of the sensing units (e.g., a node density of 1200 nodes / mm²). For example, the strain gradient in the X direction is 148 μm, the strain gradient in the Y direction is 118 μm, and the strain gradient in the Z direction is 98 μm. Ultimately, a three-dimensional strain gradient tensor is generated, providing high-precision data support for subsequent analysis.

[0177] Here is a specific example:

[0178] During the critical period of bone integration 3-6 weeks after oral implant surgery, a distributed network was used to collect multi-axial strain signals (150 μm in the X direction, 120 μm in the Y direction, and 100 μm in the Z direction) from the target contact surface based on the preset collagen fiber density gradient distribution (1200 fibers / mm² in the surface layer and 800 fibers / mm² in the deep layer). A temperature compensation algorithm was used to eliminate ambient temperature interference (temperature fluctuation of ±2°C) and generate multi-scale strain gradient vectors (148 μm in the X direction, 118 μm in the Y direction, and 98 μm in the Z direction). The multiscale strain gradient vectors were geometrically aligned with the pre-determined collagen fiber orientation to generate a multiscale strain gradient distribution matrix (with a 0.5 kPa occlusal load component). Based on the occlusal load component and the spatial distribution parameters of the sensing elements (node density 1200 nodes / mm²), 148 μm, 118 μm, and 98 μm were used as diagonal elements in the matrix, and the values of the off-diagonal elements were calculated. Specifically, k = c × occlusal load, where c = 0.1, which can be determined based on material properties and experimental data. For a 0.5 kPa occlusal load, k = 0.05 kPa. f = node density / 1000.

[0179] The off-diagonal elements can be calculated as follows:

[0180] ;

[0181] in, 148μm, 118μm, It is 98μm.

[0182] When generating the element values on the off-diagonal line, a 3×3 matrix is obtained, which is a three-dimensional strain gradient tensor. Therefore, the embodiment of the present application can reconstruct the strain gradient tensor of the target contact surface in three-dimensional space, providing a scientific basis for postoperative monitoring and early warning.

[0183] This three-step approach forms a complete risk assessment and early warning system, effectively preventing complications such as peri-implantitis caused by delayed mucosal elasticity recovery, significantly improving the success rate of implant surgery and the quality of patient recovery, and ensuring optimal treatment outcomes. This embodiment of the application introduces multi-scale strain gradient vectors and geometric space alignment technology to provide a more accurate description of strain distribution, thereby better reflecting the changes in the mechanical behavior of mucosal tissue at different healing stages and improving the reliability and accuracy of the overall monitoring system.

[0184] Figure 2 The present invention provides a schematic diagram of the structure of an intelligent prediction and early warning system for oral implant risks, as shown in FIG. Figure 2 As shown, the system includes:

[0185] Acquisition module 21 is used to obtain the mechanical response data of the oral mucosal tissue under occlusal load and the three-dimensional strain gradient tensor of the target contact surface, and establish a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue throughout the healing cycle in combination with a preset stress-strain analysis model. The target contact surface is the contact surface between the implant and the oral mucosal tissue;

[0186] an analysis module 22 for performing a coupled analysis on the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus to generate multi-level relaxation correlation parameters, and analyzing the stress conduction path offset of the target contact surface by combining the multi-level relaxation correlation parameters with a lightweight network;

[0187] The generating module 23 is configured to generate a hierarchical early warning instruction according to the spatiotemporal correlation between the stress conduction path offset and the preset healing stage biomechanical threshold.

[0188] Figure 2 The intelligent prediction and early warning system for oral implant risks can be implemented Figure 1 The implementation principles and technical effects of the intelligent prediction and early warning method for oral implant risks described in the illustrated embodiment will not be elaborated on here. The specific manner in which each module and unit performs operations in the intelligent prediction and early warning system for oral implant risks in the above embodiment has been described in detail in the embodiments of the method and will not be elaborated on here.

[0189] In one possible design, Figure 2 The intelligent prediction and early warning system for oral implant risks in the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0190] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0191] The processing component 32 is as follows Figure 1 The embodiment provides an intelligent prediction and early warning method for oral implant risks.

[0192] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0193] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0194] Of course, a computing device may also include other components, such as input / output interfaces, display components, and communication components. The input / output interfaces provide an interface between the processing component and peripheral interface modules, which may be output devices, input devices, etc. The communication components are configured to facilitate wired or wireless communication between the computing device and other devices.

[0195] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0196] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment provides an intelligent prediction and early warning method for oral implant risks.

[0197] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0198] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0199] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent prediction and early warning method for oral implant risks, characterized in that: Smart devices used for monitoring abnormal wound healing during implant surgery include: Acquiring mechanical response data of oral mucosal tissue under occlusal load and a three-dimensional strain gradient tensor of a target contact surface, and establishing a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue throughout the healing cycle in combination with a preset stress-strain analysis model, wherein the target contact surface is the contact surface between the implant and the oral mucosal tissue; Couple the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus to generate multi-level relaxation correlation parameters, and analyze the stress conduction path offset of the target contact surface by combining the multi-level relaxation correlation parameters with a lightweight network; generating hierarchical early warning instructions based on the spatiotemporal correlation between the stress conduction path offset and the preset healing stage biomechanical threshold; The coupling analysis of the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus to generate multi-level relaxation correlation parameters, and analyzing the stress conduction path offset of the target contact surface by combining the multi-level relaxation correlation parameters with a lightweight network, includes: Based on the target dynamic attenuation curve of the elastic modulus, a dynamic matching framework is established in combination with the correlation between the stress distribution pattern in the preset stress-strain analysis model and the preset collagen fiber orientation; Determining the real-time relaxation fluctuation characteristics of the target contact surface under the occlusal load based on the three-dimensional strain gradient tensor, removing environmental interference from the real-time relaxation fluctuation characteristics, and combining the dynamic matching framework and the density gradient distribution of the distributed network to quantify the spatiotemporal superposition effect of the occlusal load amplitude during the critical period of bone integration in different healing stages of the entire healing process, thereby generating multi-level relaxation correlation parameters. The distributed network is the physical network architecture of collagen fibers in the oral mucosal tissue. The multi-level relaxation correlation parameters are used to reflect the changes in the mechanical properties of the oral mucosal tissue throughout the healing cycle. Based on the multi-level relaxation association parameters, a stress conduction path model of the target contact surface is constructed. The dynamic matching framework and the stress conduction path model are coupled and analyzed using a lightweight network, and the stress conduction path offset is output. The lightweight network adopts a deep separable convolution architecture.

2. The method according to claim 1, characterized in that The method of constructing a stress conduction path model of the target contact surface based on the multi-level relaxation association parameters, coupling the dynamic matching framework and the stress conduction path model using a lightweight network, and outputting a stress conduction path offset includes: Based on the multi-level relaxation correlation parameters, the sensing units of the distributed network are used as nodes and the stress distribution pattern is used as an edge to establish a stress conduction path model of the target contact surface; Using a lightweight network, the stress conduction path model is decomposed into spatial dimension features and channel dimension features. The spatial dimension features are optimized in combination with the bite load direction of the three-dimensional strain gradient tensor, and the channel dimension features are optimized in combination with the dynamic attenuation phase of the target dynamic attenuation curve to obtain processed spatial dimension features and processed channel dimension features. Comparing the strain distribution data in the processed spatial dimensional features with the node stress distribution in the stress conduction path model point by point to generate a node stress conduction response sequence, comparing the attenuation rate in the processed channel dimensional features with the edge stress distribution in the stress conduction path model edge by edge to generate an edge stress conduction response sequence, and combining the node stress conduction response sequence and the edge stress conduction response sequence to generate a multi-scale stress conduction response sequence; Based on the comparison of the attenuation rate of the compression relaxation component and the attenuation rate of the shear relaxation component in the dynamic matching framework, an attenuation rate difference is obtained. According to the attenuation rate difference, a path offset accumulation calculation is performed on the multi-scale stress conduction response sequence to output the stress conduction path offset.

3. The method according to claim 2, characterized in that The step of performing path deviation accumulation calculation on the multi-scale stress conduction response sequence according to the attenuation rate difference and outputting the stress conduction path deviation comprises: establishing a relaxation decay rate difference model of the target contact surface according to the decay rate difference; According to the spatiotemporal superposition effect of the multi-scale stress conduction response sequence, the stress distribution pattern in the preset stress-strain analysis model and the density gradient distribution of the preset collagen fiber distributed network are superimposed in the relaxation attenuation rate difference model to generate a cumulative path deviation vector; Based on the real-time relaxation fluctuation characteristics, the accumulated path offset vector is subjected to phase synchronization filtering to eliminate ambient temperature interference of the distributed network, and a corrected accumulated path offset vector is output; Performing dynamic weighted fusion on the corrected cumulative path offset vector, the compression relaxation component and the shear relaxation component in the dynamic matching framework, and adjusting the weighting coefficient according to the dynamic attenuation phase during the dynamic weighted fusion process to generate a path offset dynamic correlation parameter; Based on the dynamic interactive coupling results of the path offset dynamic association parameters and the multi-level relaxation association parameters, the weights of the nodes and edges of the stress conduction path model are iteratively updated through the distributed network to output the stress conduction path offset.

4. The method according to claim 1, wherein The method determines the real-time relaxation fluctuation characteristics of the target contact surface under the occlusal load based on the three-dimensional strain gradient tensor, removes environmental interference in the real-time relaxation fluctuation characteristics, combines the dynamic matching framework and the density gradient distribution of the distributed network, quantifies the spatiotemporal superposition effect of the occlusal load amplitude during the critical period of bone integration in different healing stages of the entire healing process, and generates multi-level relaxation correlation parameters, including: establishing a real-time relaxation fluctuation model of the oral mucosal tissue under the occlusal load based on the three-dimensional strain gradient tensor, and extracting the real-time relaxation fluctuation characteristics of the target contact surface under the occlusal load from the real-time relaxation fluctuation model; Eliminating ambient temperature interference from the real-time relaxation fluctuation characteristics using a temperature compensation algorithm to generate a corrected real-time relaxation fluctuation characteristic, and inputting the corrected real-time relaxation fluctuation characteristic into the dynamic matching framework to generate a composite load relaxation fluctuation parameter; quantifying the spatiotemporal superposition effect of the composite load relaxation fluctuation parameter according to the density gradient distribution of the distributed network, and generating a time window accumulation function; The time window cumulative function is dynamically coupled with the density gradient distribution in units of time to obtain a dynamic coupling result. The dynamic coupling result is matched with the target dynamic attenuation curve to generate multi-level relaxation association parameters, wherein the multi-level relaxation association parameters include a decay rate parameter of a compression relaxation component and a decay rate parameter of a shear relaxation component. The decay rate parameter of each component is the product of the corresponding decay rate and the corresponding weight.

5. The method according to claim 1, wherein The method acquires the mechanical response data of the oral mucosal tissue under occlusal load and the three-dimensional strain gradient tensor of the target contact surface, and establishes a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue throughout the healing cycle in combination with a preset stress-strain analysis model, wherein the target contact surface is the contact surface between the implant and the oral mucosal tissue, including: Acquiring mechanical response data of oral mucosal tissue under occlusal load, extracting a compression relaxation component, a shear relaxation component, and an elastic modulus from the mechanical response data, analyzing the decay rates of the compression relaxation component and the shear relaxation component over time using a preset stress-strain analysis model, and generating a decay rate weight; Based on the attenuation rate weights, a multi-scale attenuation coefficient matrix is established, wherein the multi-scale attenuation coefficient matrix is used to describe the attenuation rate weight distribution of the compression relaxation component and the shear relaxation component in different healing stages of the entire healing process, wherein the healing stage includes the critical period of bone integration; The elastic modulus is divided into dynamic attenuation stages according to the multi-scale attenuation coefficient matrix, and the dynamic attenuation stages and the attenuation trajectory of the combined compression relaxation component and the shear relaxation component are hierarchically associated to generate a time-varying gradient parameter of the oral mucosal tissue throughout the healing process. The attenuation trajectory refers to a curve of the compression relaxation component and the shear relaxation component changing over time during the healing cycle, reflecting the dynamic attenuation process of the mechanical properties of the oral mucosal tissue; The three-dimensional strain gradient tensor of the target contact surface is obtained in real time through a distributed network based on the preset correlation relationship of the collagen fiber orientation; Based on the three-dimensional strain gradient tensor and in combination with the time-varying gradient parameter, a multi-scale stress relaxation iterative correction is performed on the stress distribution pattern in the preset stress-strain analysis model to generate an initial dynamic attenuation curve of the elastic modulus; By utilizing the biomechanical property constraints of the entire healing cycle of oral mucosal tissue, the decay rates of the compression relaxation component and the shear relaxation component in the initial dynamic decay curve of the elastic modulus are phase-synchronized to match the spatiotemporal distribution of the decay rate with the occlusal load amplitude during the critical period of osseointegration, thereby obtaining the decay trajectories of the compression relaxation component and the shear relaxation component; The compression relaxation component attenuation trajectory, the shear relaxation component attenuation trajectory, and the weight of the multi-scale attenuation coefficient matrix are dynamically superimposed to generate a target dynamic attenuation curve of the elastic modulus.

6. The method according to claim 5, characterized in that The method of obtaining the three-dimensional strain gradient tensor of the target contact surface in real time based on the preset correlation relationship of the collagen fiber orientation through the distributed network includes: According to the preset density gradient distribution of the distributed network of collagen fibers, multi-axial strain signals of the target contact surface are collected, and the multi-axial strain signals are eliminated from ambient temperature interference using a temperature compensation algorithm to generate a multi-scale strain gradient vector; Aligning the multi-scale strain gradient vector with a preset collagen fiber orientation in geometric space to generate a multi-scale strain gradient distribution matrix, wherein the multi-scale strain gradient distribution matrix includes an occlusal load component; Based on the bite load component and in combination with the spatial distribution parameters of the sensing unit, the strain gradient tensor of the target contact surface in three-dimensional space is reconstructed to generate a three-dimensional strain gradient tensor.

7. An intelligent prediction and early warning system for oral implant risks, characterized by: include: an acquisition module, configured to acquire the mechanical response data of the oral mucosal tissue under occlusal load and the three-dimensional strain gradient tensor of the target contact surface, and establish a target dynamic attenuation curve of the elastic modulus of the oral mucosal tissue throughout the healing cycle in combination with a preset stress-strain analysis model, wherein the target contact surface is the contact surface between the implant and the oral mucosal tissue; an analysis module for coupling analysis of the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus to generate multi-level relaxation correlation parameters, and analyzing the stress conduction path offset of the target contact surface by combining the multi-level relaxation correlation parameters with a lightweight network; a generation module, configured to generate hierarchical early warning instructions based on the spatiotemporal correlation between the stress conduction path offset and a preset healing stage biomechanical threshold; The coupling analysis of the three-dimensional strain gradient tensor and the target dynamic attenuation curve of the elastic modulus to generate multi-level relaxation correlation parameters, and analyzing the stress conduction path offset of the target contact surface by combining the multi-level relaxation correlation parameters with a lightweight network, includes: Based on the target dynamic attenuation curve of the elastic modulus, a dynamic matching framework is established in combination with the correlation between the stress distribution pattern in the preset stress-strain analysis model and the preset collagen fiber orientation; Determining the real-time relaxation fluctuation characteristics of the target contact surface under the occlusal load based on the three-dimensional strain gradient tensor, removing environmental interference from the real-time relaxation fluctuation characteristics, and combining the dynamic matching framework and the density gradient distribution of the distributed network to quantify the spatiotemporal superposition effect of the occlusal load amplitude during the critical period of bone integration in different healing stages of the entire healing process, thereby generating multi-level relaxation correlation parameters. The distributed network is the physical network architecture of collagen fibers in the oral mucosal tissue. The multi-level relaxation correlation parameters are used to reflect the changes in the mechanical properties of the oral mucosal tissue throughout the healing cycle. Based on the multi-level relaxation association parameters, a stress conduction path model of the target contact surface is constructed. The dynamic matching framework and the stress conduction path model are coupled and analyzed using a lightweight network, and the stress conduction path offset is output. The lightweight network adopts a deep separable convolution architecture.

8. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an intelligent prediction and early warning method for oral implant risks as described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, an intelligent prediction and early warning method for oral implant risks as described in any one of claims 1 to 6 is implemented.

Images