Skin state evaluation method and system based on torque trigger data

Abstract

The application discloses a skin state evaluation method and system based on torque trigger data, and belongs to the technical field of skin state detection. The method comprises the following steps: acquiring skin part information and searching a database to obtain scene adaptation parameters; establishing a body surface coordinate system and collecting contact pressure data and probe attitude data; controlling the probe to move and performing dynamic torque balance regulation and control to obtain dynamic torque balance regulation and control parameters; performing feature matching based on the parameters and a reference model to obtain a feature trigger signal; recording a sliding distance and collecting synchronous data in response to the signal to obtain a complete torque displacement correlation data set; extracting skin biomechanics feature parameters; and obtaining skin state classification labels through fusion calculation. The application adopts a scheme of loading a reference model according to a skin part, performing dynamic torque balance regulation and control in sliding, and collecting a termination point based on torque curve feature trigger data, and can realize adaptive, accurate and objective evaluation of the skin biomechanics state.

Description

Technical Field

[0001] This invention relates to the field of skin condition detection technology, and in particular to a skin condition assessment method and system based on torque triggering data. Background Technology

[0002] As the largest organ in the human body, the health of the skin is a key focus in clinical medicine and skincare. The biomechanical properties of the skin, such as elasticity, viscosity, firmness, and friction characteristics, are crucial objective indicators characterizing its physiological state, aging level, and moisture content. Therefore, developing diagnostic devices for accurately and non-invasively measuring skin biomechanical properties is of great significance for the auxiliary diagnosis of skin diseases, the evaluation of skincare product efficacy, and the formulation of personalized care plans.

[0003] In existing technologies, the measurement of skin biomechanical properties typically employs various probe-based instruments. These instruments generally operate by controlling the probe's contact with the skin surface and applying specific mechanical stimulation according to a pre-set program, such as constant pressure, a preset sliding speed, or a fixed sliding distance. During this process, sensors record skin response data, such as friction, deformation, or recovery time, and the skin's condition is then evaluated through the analysis of these single or a few parameters.

[0004] However, existing technical solutions have some inherent drawbacks. First, using a fixed motion procedure lacks real-time feedback and adjustment of the testing process, making it unable to adapt to the differences in skin characteristics among different individuals and body parts. This can easily lead to distorted measurement data due to improper contact pressure or poor sliding paths. Second, the starting and ending points of the measurement are usually defined by fixed time or displacement, which is rather arbitrary and fails to accurately correspond to the skin's inherent mechanical response stages, thus reducing the comparability of the data and the reliability of the results. Finally, the assessment of skin condition often relies on a few parameters such as the endpoint value or average value, ignoring the rich information contained in the dynamic response curve. This results in a somewhat one-sided assessment conclusion that fails to fully reflect the complex viscoelastic properties of the skin. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a skin condition assessment method and system based on torque-triggered data. The method employs a technical solution that loads a reference model based on the skin location, performs dynamic torque balance control during sliding, and triggers the data acquisition termination point based on torque curve characteristics. This approach enables adaptive, accurate, and objective assessment of the skin's biomechanical state.

[0006] The above objectives can be achieved through the following approach: A skin condition assessment method based on torque-triggered data includes: acquiring information about the skin site to be assessed; retrieving a scene pattern database to load the corresponding assessment scene parameter set and skin biomechanical reference model to obtain scene adaptation parameters; establishing a body surface coordinate system on the skin site to be assessed according to the scene adaptation parameters, and controlling the assessment probe to make initial contact with the skin surface to collect pressure and posture information to obtain contact pressure data and probe posture data; controlling the movement of the assessment probe according to the contact pressure data and the probe posture data, and performing adaptive sliding control and dynamic torque balance regulation during the movement to obtain dynamic torque balance regulation parameters; performing torque curve feature matching based on the dynamic torque balance regulation parameters and the skin biomechanical reference model to obtain a feature trigger signal; terminating the sliding in response to the feature trigger signal and recording the sliding distance from the sliding start point, while simultaneously collecting data from the sliding start point to the feature trigger point to obtain a complete torque displacement association dataset; performing feature extraction based on the sliding distance and the complete torque displacement association dataset to obtain skin biomechanical feature parameters; inputting the skin biomechanical feature parameters into the condition assessment model for weighted fusion and mapping calculation, and comparing them with a classification threshold to obtain a skin condition classification label.

[0007] Optionally, obtaining the scene adaptation parameters includes: matching and selecting an evaluation scene mode in the scene mode database based on the skin location information; loading associated initial pressure parameters, expected sliding direction parameters, and a skin biomechanical reference model from the evaluation scene mode, wherein the biomechanical reference model includes an ideal resistance torque displacement characteristic curve; and combining the initial pressure parameters, the expected sliding direction parameters, and the skin biomechanical reference model to obtain the scene adaptation parameters.

[0008] Optionally, obtaining the contact pressure data and probe attitude data includes: parsing the expected sliding direction parameter in the scene adaptation parameters, defining the direction consistent with the expected sliding direction as the horizontal axis of the body surface coordinate system, defining the normal feed axis of the evaluation probe as the vertical axis of the body surface coordinate system, and constructing the body surface coordinate system; extracting the initial pressure parameter in the scene adaptation parameters as a preload threshold, controlling the evaluation probe to feed along the vertical axis of the body surface coordinate system to the skin surface until the real-time value fed back by the pressure sensor reaches the preload threshold and enters the steady-state range; reading the real-time pressure value after entering the steady-state range as contact pressure data, and collecting the tilt angle data of the probe relative to the vertical axis of the body surface coordinate system through the attitude sensor set on the evaluation probe as probe attitude data.

[0009] Optionally, obtaining the dynamic torque balance control parameters includes: calculating the path following deviation based on the probe posture data and the virtual reference path established based on the body surface coordinate system; monitoring the torque change rate data of the evaluation probe during the sliding process in real time, comparing it with the ideal torque change trend defined in the skin biomechanical reference model, and calculating the pressure adjustment amount or direction correction amount in combination with the path following deviation to obtain the adjusted contact pressure data or the guiding signal used to correct the sliding direction; and combining the torque change rate data, the path following deviation, and the adjusted contact pressure data or the guiding signal to obtain the dynamic torque balance control parameters.

[0010] Optionally, obtaining the adjusted contact pressure data or the guiding signal for correcting the sliding direction includes: when the torque change rate data exceeds the upper limit threshold corresponding to the ideal torque change trend, generating a pressure reduction control command or generating a guiding command prompting an increase in path offset; when the torque change rate data is lower than the lower limit threshold corresponding to the ideal torque change trend, generating a pressure increase control command or generating a guiding command prompting correction to the virtual reference path.

[0011] Optionally, obtaining the characteristic trigger signal includes: extracting real-time torque change rate data from the dynamic torque balance control parameters, and determining whether it enters and remains in the range near the zero value used to characterize the stable state, thereby obtaining the equilibrium maintenance state; if the duration of the equilibrium maintenance state reaches the stability duration threshold used to determine the equilibrium state, then it is determined that the torque curve characteristic trigger condition is met, and the characteristic trigger signal is obtained.

[0012] Optionally, obtaining the complete torque-displacement correlation dataset includes: acquiring the pulse count value of the displacement recording unit used to monitor the movement of the evaluation probe at the moment the feature trigger signal is generated; calculating the algebraic difference between the pulse count value and the initial count value at the sliding start point to obtain the sliding distance; retrieving the torque sampling sequence and displacement sampling sequence recorded from the sliding start point as the original sampling data; and using the sliding distance to perform interval truncation and spatiotemporal alignment on the original sampling data based on the displacement endpoint to generate the complete torque-displacement correlation dataset.

[0013] Optionally, obtaining the skin biomechanical characteristic parameters includes: performing numerical integration calculation on the complete torque-displacement correlation dataset to obtain the integral area under the torque-displacement curve as a first characteristic parameter; extracting the peak value of the rate of change before the generation of the feature trigger signal from the complete torque-displacement correlation dataset to obtain the peak value of the torque rate of change as a second characteristic parameter; extracting the sliding distance as a third characteristic parameter, and combining it with the first characteristic parameter and the second characteristic parameter to obtain the skin biomechanical characteristic parameters.

[0014] Optionally, obtaining the skin condition classification label includes: inputting the skin biomechanical feature parameters into a condition assessment model trained with historical data for processing to obtain a feature mapping result; performing weighted fusion of the feature mapping result through the condition assessment model to obtain a skin condition index; and comparing the skin condition index with a classification threshold to obtain a skin condition classification label.

[0015] Based on the same inventive concept, this invention also provides a skin condition assessment system based on torque-triggered data, comprising: a scene parameter matching module, used to acquire information of the skin part to be assessed, retrieve a scene pattern database to load the corresponding assessment scene parameter set and skin biomechanical reference model, and obtain scene adaptation parameters; a posture pressure sensing module, used to establish a body surface coordinate system on the skin part to be assessed according to the scene adaptation parameters, and control the assessment probe to make initial contact with the skin surface to collect pressure and posture information, and obtain contact pressure data and probe posture data; and a dynamic control and regulation module, used to control the movement of the assessment probe according to the contact pressure data and the probe posture data, and perform adaptive sliding control and dynamic torque balance regulation during the movement, and obtain dynamic torque... The system includes: a torque balance control parameter; a feature signal triggering module, used to perform torque curve feature matching based on the dynamic torque balance control parameter and the skin biomechanical reference model to obtain a feature triggering signal; an associated data acquisition module, used to terminate sliding in response to the feature triggering signal and record the sliding distance from the sliding start point, while simultaneously collecting data from the sliding start point to the feature triggering point to obtain a complete torque-displacement associated dataset; a mechanical feature extraction module, used to extract features based on the sliding distance and the complete torque-displacement associated dataset to obtain skin biomechanical feature parameters; and a state decision classification module, used to input the skin biomechanical feature parameters into a state assessment model for weighted fusion and mapping calculation, and compare them with a classification threshold to obtain a skin state classification label.

[0016] Compared with the prior art, the present invention has the following advantages: This invention introduces a scene pattern database and a skin biomechanical reference model to conduct standardized assessments of the characteristics of different skin areas. Combined with an automated execution process, it eliminates interference from differences in human operation and subjective judgment, thereby improving the objectivity and repeatability of the assessment results.

[0017] This invention proposes an adaptive sliding control and dynamic torque balance regulation strategy, which can monitor and adjust the mechanical interaction state between the probe and the skin in real time during the evaluation process, keeping it within a range close to the ideal model. This compensates for the influence of variables such as microscopic unevenness of the skin surface, ensuring the quality and high fidelity of the original data acquisition.

[0018] This invention employs a triggering mechanism based on torque curve characteristics to determine the termination point of data acquisition, ensuring that each measurement fully captures the critical process with clear biomechanical significance from dynamic response to steady-state transition. This results in the extracted characteristic parameters, such as sliding distance, having stronger intrinsic consistency and physical meaning, thereby improving the accuracy of the assessment.

[0019] This invention extracts multi-dimensional feature parameters such as integral area, peak torque change rate, and sliding distance, and inputs them into a state assessment model for weighted fusion and comprehensive decision-making. This enables a comprehensive characterization of the skin's mechanical properties from multiple perspectives, including energy dissipation, instantaneous stiffness, and extensibility, achieving in-depth analysis and intelligent classification of skin conditions, resulting in more comprehensive and reliable assessment results.

[0020] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic flowchart of a skin condition assessment method based on torque-triggered data according to an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram of the characteristic curves of the skin biomechanical reference model according to an embodiment of the present invention.

[0024] Figure 3 This is a radar diagram of multidimensional skin biomechanical characteristic parameters according to an embodiment of the present invention.

[0025] Figure 4 This is a schematic diagram of a skin condition assessment system based on torque-triggered data according to an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Reference Figure 1 One embodiment of the present invention proposes a skin condition assessment method based on torque-triggered data. The method employs a technical solution that loads a reference model on the skin site, performs dynamic torque balance control during sliding, and triggers the data acquisition termination point based on the characteristics of the torque curve. This method can achieve adaptive, accurate, and objective assessment of the skin's biomechanical state.

[0028] The method described in this embodiment specifically includes: S1. Obtain information about the skin area to be evaluated, retrieve the scene pattern database to load the corresponding evaluation scene parameter set and skin biomechanical reference model, and obtain scene adaptation parameters. Optionally, the obtained scene adaptation parameters include: Based on the skin area information, match and select an evaluation scene mode from the scene mode database; Load the associated initial pressure parameters, expected sliding direction parameters, and skin biomechanical reference model from the evaluation scenario mode, wherein the biomechanical reference model includes an ideal drag torque displacement characteristic curve; By combining the initial pressure parameters, the expected sliding direction parameters, and the skin biomechanical reference model, scene adaptation parameters are obtained.

[0029] Specifically, the process begins with receiving specific skin location information, which is a string or code used as a query index, such as "inner forearm" or "cheekbone area of ​​the face". Using this information as a key field, a matching search is performed in a built-in scene pattern database to locate and select the evaluation scene pattern that perfectly matches it. This scene pattern database is a pre-built structured dataset in which each evaluation scene pattern encapsulates a measurement configuration optimized for specific body part skin characteristics, such as curvature, thickness, and elasticity.

[0030] After successfully selecting the evaluation scenario mode, a set of interrelated core parameters are loaded from that mode. First is the initial pressure parameter, a defined target value for the normal force, typically set between 0.5 and 2.0 Newtons, used to ensure the evaluation probe establishes stable and consistent initial contact with the skin before sliding begins. Second is the expected sliding direction parameter, which defines the ideal movement path of the probe in the body surface coordinate system, usually represented as a two-dimensional or three-dimensional vector, guiding subsequent sliding control. Most importantly, a skin biomechanical reference model associated with this scenario is loaded. This model specifically uses an exponentially modified nonlinear viscoelastic constitutive equation to construct the ideal drag torque displacement characteristic curve. The specific mathematical expression is as follows: , in, The ideal resistance torque is the equivalent torque generated by the probe when it overcomes the sliding resistance of the skin, and d is the current sliding displacement. It serves as the reference frictional torque constant, used to characterize the fundamental torque transformed from pure surface friction when no significant biomechanical deformation occurs in the initial stage of sliding; is the elastic modulus of the drag torque, used to characterize the gain of the skin's initial softness in the small deformation stage corresponding to the torque dimension; The strain hardening index characterizes the rate of nonlinear stiffness increase caused by the gradual tightening and directional alignment of collagen fiber bundles in the dermis with increasing displacement. For example... Figure 2 As shown in the figure, the curve of the loaded biomechanical reference model is displayed after selecting the evaluation scenario mode.

[0031] The initial pressure parameters, expected sliding direction parameters, and skin biomechanical reference model are combined and integrated into a single data structure or parameter set. To ensure that this heterogeneous data can be uniformly accessed by the control system, this combination process logically constructs an encapsulated container, such as a structure or key-value pair object. Space is allocated in memory for this container, mapping pressure values ​​to a "force control threshold" field, direction vectors to a "path planning" field, and serializing and storing the coefficients of the reference model in a "baseline model" field. The integrated data structure constitutes the scene adaptation parameters.

[0032] For example, the user inputs "cheekbone area" as the test site through the interactive interface. The system uses this string as the query key to search the database and matches the evaluation scenario mode with ID FACE_ZYG_V2. The system then loads the associated parameters for this mode: the initial pressure parameter is locked at 1.0 Newtons; the expected sliding direction parameter is set as a horizontal vector (1,0,0) along the zygomatic arch towards the ear; and the skin biomechanical reference model loading coefficients are used. Newton rice Newton rice To verify the correctness of the model loading, the system internally performed pre-calculations when the sliding displacement... The ideal drag torque at a given time is substituted into the formula. The calculation process is as follows: Newton-meters. After confirming that this value is within the normal resistance range of human facial skin, the system encapsulates all the above parameters and coefficients into a structure named SceneConfig, completing the generation of scene adaptation parameters. This method, by loading location-specific scene parameters and biomechanical models, ensures that the measurement process strictly follows the physiological characteristics of different anatomical sites, avoiding measurement errors or safety risks caused by using general parameters, thereby improving the adaptability and accuracy of the evaluation system in complex human body surface applications.

[0033] S2. Establish a body surface coordinate system on the skin area to be evaluated according to the scene adaptation parameters, and control the evaluation probe to make initial contact with the skin surface to collect pressure and posture information, and obtain contact pressure data and probe posture data. Optionally, obtaining the contact pressure data and probe attitude data includes: The expected sliding direction parameter in the scene adaptation parameters is analyzed, and the direction consistent with the expected sliding direction is defined as the horizontal axis of the body surface coordinate system, and the normal feed axis of the evaluation probe is defined as the vertical axis of the body surface coordinate system to construct the body surface coordinate system. The initial pressure parameter in the scene adaptation parameters is extracted as the preload threshold, and the evaluation probe is controlled to advance along the longitudinal axis of the body surface coordinate system to the skin surface until the real-time value fed back by the pressure sensor reaches the preload threshold and enters the steady state range. The real-time pressure value after entering the steady-state range is read as the contact pressure data, and the tilt angle data of the probe relative to the longitudinal axis of the body surface coordinate system is collected by the attitude sensor set on the evaluation probe as the probe attitude data.

[0034] Specifically, the scene adaptation parameters are analyzed, and the expected sliding direction parameters are extracted from them. This vector direction is defined as the horizontal axis of the body surface coordinate system on the skin area to be evaluated, i.e., the main direction of sliding. Simultaneously, the mechanical motion axis of the evaluation probe performing normal feed is defined as the vertical axis of the body surface coordinate system. The body surface coordinate system is a temporary Cartesian coordinate system established locally on the skin for a single measurement. Its function is to provide a unified reference framework for all subsequent motion control and data acquisition, ensuring the consistency of displacement and attitude data.

[0035] Initial pressure parameters are extracted from the scene adaptation parameters and used as a preload threshold. The control system drives the evaluation probe to advance towards the skin surface at a controlled speed along the defined longitudinal axis of the body surface coordinate system. During this process, a pressure sensor integrated into the probe tip provides high-frequency feedback of real-time pressure values. When this real-time value first reaches the preload threshold, such as 1.0 Newtons, the control system enters a steady-state holding phase. At this point, the probe position is continuously fine-tuned to maintain the pressure value within a very small fluctuation range, which is the steady-state interval. This interval is defined as within ±5% of the target pressure value and lasts for more than a stable duration, such as 200 milliseconds.

[0036] Once the pressure is determined to have entered and maintained within the steady-state range, it signifies that initial contact has been stably established. At this point, data acquisition is immediately initiated. First, the average real-time pressure value within the steady-state range is read and recorded as the contact pressure data for this evaluation. This data point serves as the baseline pressure for subsequent dynamic control. Simultaneously, the current attitude information of the probe is acquired using an attitude sensor mounted on the evaluation probe, typically an inertial measurement unit integrating an accelerometer and a gyroscope. By analyzing the sensor data, the deviation angle of the probe's actual axis relative to the ideal longitudinal axis of the body surface coordinate system is calculated; this angle data constitutes the probe attitude data.

[0037] For example, the system analyzes the scene adaptation parameters, establishes a body surface coordinate system with the Z-axis along the opposite direction of gravity, and controls the probe to feed along the Z-axis at a speed of 0.5 mm / s. The preload threshold is set to 1.0 Newtons. When the pressure sensor feedback value first reaches 1.0 Newtons, the system enters steady-state hold control. During the next 200 ms steady-state interval, the pressure values ​​of five key sampling points are collected as 0.98 Newtons, 1.01 Newtons, 0.99 Newtons, 1.02 Newtons, and 1.00 Newtons, respectively, and the system calculates their arithmetic mean. Newton's time was used as the final contact pressure data. Simultaneously, the probe's built-in inertial measurement unit (IMU) acquired the current attitude. Ideally, the Z-axis corresponding gravitational acceleration component should be 1.0g, while the measured Z-axis component is 0.996g. The tilt angle was calculated using the inverse cosine function. The radian value, approximately 5.1 degrees, is recorded as probe posture data and used to subsequently compensate for mechanical component errors caused by the probe not being perpendicular to the surface. This method establishes a precise initial steady-state and posture reference before dynamic testing begins, effectively filtering out uncertainties in initial conditions caused by differences in operation techniques or minor undulations in the skin surface, ensuring that the subsequently acquired biomechanical data has a unified zero-point reference and repeatability.

[0038] S3. Control the movement of the evaluation probe according to the contact pressure data and the probe posture data, and perform adaptive sliding control and dynamic torque balance regulation during the movement to obtain dynamic torque balance regulation parameters. Optionally, the obtained dynamic torque balance control parameters include: The path following deviation is calculated based on the probe posture data and the virtual reference path established based on the body surface coordinate system. The torque change rate data of the evaluation probe during the sliding process is monitored in real time, compared with the ideal torque change trend defined in the skin biomechanical reference model, and the pressure adjustment amount or direction correction amount is calculated in combination with the path following deviation to obtain the adjusted contact pressure data or the guiding signal used to correct the sliding direction. By combining the torque change rate data, the path following deviation, and the adjusted contact pressure data or the guiding signal, dynamic torque balance control parameters are obtained.

[0039] Specifically, this process is executed in a high-frequency control loop, updated for example every 10 to 20 milliseconds. Based on previously acquired probe attitude data and combined with a virtual reference path established using the body surface coordinate system, the path following deviation is calculated in real time. The virtual reference path is the geometric realization of the expected sliding direction parameter in the scene adaptation parameters in the body surface coordinate system, typically a straight line. The path following deviation is a quantitative indicator characterizing the degree to which the probe's current position and attitude deviate from this ideal path.

[0040] The torque sensor signal integrated into the evaluation probe is monitored and processed in real time, and the instantaneous torque change rate data is obtained through differential calculation. Simultaneously, the ideal torque change trend is retrieved from the skin biomechanical reference model based on the current sliding displacement or time. This is a preset baseline change rate, representing the standard response at that specific skin location. The measured torque change rate data is compared with the ideal torque change trend to calculate the torque response error. Subsequently, the core control algorithm uses the path following deviation and torque response error as dual inputs to calculate a comprehensive control command, namely, pressure adjustment or direction correction. This calculation process can be represented by the following control law: , in, This represents the final generated control command; The torque response error, which is the difference between the measured torque change rate and the ideal torque change trend, is obtained through real-time monitoring. The path following deviation is obtained by comparing the position of the attitude sensor data with that of the virtual reference path. and These are preset weighting coefficients stored in the scene adaptation parameter set. They determine the system's sensitivity to torque and path errors and represent dimensionless adjustment gains. According to... The calculation results are used to generate adjusted contact pressure data or guide signals for correcting the sliding direction.

[0041] The core data throughout the entire control cycle—namely, real-time torque change rate data, path following deviation, and the final adjusted contact pressure data or guide signal—are synchronized and packaged in the time domain. Specifically, this is achieved by constructing a synchronized data frame with a unified timestamp. Using the current sampling time as a reference, the three types of data are filled into the payload area of ​​this data frame according to a predetermined protocol, ensuring strict alignment of mechanical data, position data, and control commands on the time axis. The synchronized data frame containing complete timing information constitutes the dynamic torque balance control parameters.

[0042] For example, at a certain moment during the sliding process The system executes a high-frequency control cycle. First, based on the attitude sensor data and the virtual reference path, the deviation of the probe's current position from the reference path by 0.5 mm is calculated. After normalization, the path following deviation is obtained. Meanwhile, the torque sensor measured the current torque change rate to be 0.015 Nm / s, while according to the skin biomechanical reference model, the ideal change rate at the current displacement should be 0.010 Nm / s. The torque response error was then calculated. Newton-meters per second. The system calls the preset weighting coefficients. and Substitute into the control law formula Perform calculations, that is The calculated result, 1.5, serves as the core value for the control command. Combined with the current timestamp and raw sensor data, it generates the dynamic torque balance control parameters for that moment, indicating the system is currently facing a state of slightly higher resistance and significant path deviation. This process achieves real-time decoupling and coordinated control of path holding and mechanical response. It not only guides the probe along the correct anatomical path but also dynamically adjusts the control strategy based on real-time skin resistance feedback, preventing data distortion caused by forced movement when the skin has non-uniform characteristics such as hardening or laxity.

[0043] Optionally, obtaining the adjusted contact pressure data or the guiding signal for correcting the sliding direction includes: When the torque change rate data exceeds the upper limit threshold corresponding to the ideal torque change trend, a pressure reduction control command or a prompt to increase path offset is generated. When the torque change rate data is lower than the lower limit threshold corresponding to the ideal torque change trend, a pressure increase control command is generated or a guidance command prompting correction to the virtual reference path is generated.

[0044] Specifically, this control logic executes in a high-frequency loop. Its core function is to determine whether the real-time torque change rate data falls within the normal range of the ideal torque change trend defined by the skin biomechanical reference model. This ideal trend is not a single curve, but a dynamic range defined by an upper and lower threshold. When the real-time monitored torque change rate data exceeds the upper threshold corresponding to this ideal torque change trend, it is determined that the interaction force between the probe and the skin is too large, potentially leading to data distortion or user discomfort. In this case, immediate intervention is implemented. A clear pressure reduction control command is generated, for example, a command to reduce the normal force by 0.1 to 0.3 Newtons, and sent to the pressure control unit of the evaluation probe. In certain operating modes, if maintaining the path is prioritized, or if pressure adjustment has reached its limit, a guiding command to increase the path offset is generated. This command may be presented to the operator in the form of visual or auditory signals, guiding them to fine-tune the sliding trajectory in the direction of less resistance.

[0045] Conversely, when the real-time monitored torque change rate data is lower than the lower threshold corresponding to the ideal torque change trend, it is determined that the contact between the probe and the skin may be insufficient, resulting in too little friction and inability to effectively collect biomechanical characteristics. In this case, a reverse control strategy is triggered. It generates a pressure increase control command, instructing the pressure control unit to appropriately increase the normal force to re-establish effective measurement contact. Alternatively, in path-first mode, a guidance command is generated to prompt correction to a virtual reference path, as deviation from the predetermined measurement area may also lead to an abnormal decrease in torque.

[0046] For example, based on the calculated control command value, the system further determines the specific execution strategy. The currently monitored torque change rate of 0.015 Nm / s exceeds the upper limit threshold of 0.012 Nm / s set for the ideal torque change trend, indicating that skin tightness is beyond expectations or friction is excessive. Based on the logic that "when the torque change rate exceeds the upper limit threshold, a pressure reduction control command is generated," the system decides to reduce the contact pressure to alleviate resistance. Let the pressure adjustment step size coefficient be... The system calculates the reduction in target pressure by adjusting the pressure using Newton's constant and the intensity of the adjustment. For simplification, a direct command can be issued to reduce the current target normal force of 1.0 Newtons by 0.1 Newtons, generating an adjusted contact pressure of 0.9 Newtons, which is then sent to the motor controller. Alternatively, if the system is in path-priority mode, a voice guidance signal, "Please slightly deflect outwards," is generated to guide the operator to correct the sliding direction to conform to the virtual reference path. This active pressure regulation and direction guidance mechanism constructs a closed-loop protection system that can immediately intervene physically when abnormally high resistance or poor contact is detected. This protects the subject's skin tissue from excessive stretching and ensures that the sensor remains within its optimal linear operating range.

[0047] S4. Based on the dynamic torque balance control parameters and the skin biomechanical reference model, torque curve feature matching is performed to obtain the feature trigger signal; Optionally, the obtained feature trigger signal includes: Real-time torque change rate data is extracted from the dynamic torque balance control parameters, and it is determined whether the torque changes into and remains in the range near the zero value used to characterize the stable state, thus obtaining the equilibrium maintenance state. If the duration of the equilibrium state reaches the stability time threshold used to determine the equilibrium state, then the torque curve characteristic triggering condition is satisfied, and a characteristic triggering signal is obtained.

[0048] Specifically, real-time torque change rate data streams are continuously extracted from the continuously generated dynamic torque balance control parameters. This data stream reflects the instantaneous rate of change of torque experienced by the evaluation probe during sliding. An internal numerical comparator continuously compares the latest torque change rate data with a preset interval near zero, used to characterize a steady state. This interval is an extremely narrow numerical range, for example, set between ±0.01 Nm / s, and its function is to identify a state where torque growth essentially stops, i.e., an equilibrium state.

[0049] When the torque change rate data first enters this near-zero range, the internal status flag is set, indicating the entry into a potential equilibrium state, and an internal timer is started simultaneously. This state is defined as the beginning of the equilibrium maintenance state. In each subsequent monitoring cycle, it continues to check whether the torque change rate remains within this range. If the data point remains within the range, the timer continues to accumulate time; once any data point exceeds the range, the status flag is immediately reset and the timer is cleared, indicating that the equilibrium state has been broken. The current value of the timer is compared in parallel with a preset stabilization duration threshold. This threshold is typically set between 50 and 150 milliseconds, and its engineering significance is to effectively filter transient noise, ensuring that the system captures a true, continuous equilibrium state, rather than random signal fluctuations. Once the accumulated time of the timer reaches or exceeds this stabilization duration threshold, it is immediately determined that the torque curve characteristic trigger condition has been met. At this time, a digital characteristic trigger signal is generated. This signal, as a high-priority event interrupt, is immediately broadcast to the system's motion control and data acquisition modules as the final instruction to terminate sliding and perform data interception.

[0050] For example, as the probe continues to slide, the skin is gradually stretched to its limit, and the increase in resistance slows down. The system continuously monitors the real-time torque change rate data extracted from the control parameters. The interval near zero, used to characterize the steady state, is set to [-0.001, +0.001] Nm / s. At time point... At that time, the rate of change was 0.003, and no trigger was triggered; At a certain point, the rate of change drops to 0.0008, falling within this range, and the system's internal timer starts. In subsequent consecutive sampling points, the rate of change remains between 0.0005 and 0.0009. When the timer's accumulated time reaches the preset stable duration threshold of 100 milliseconds, the system determines that the skin has entered quasi-static equilibrium. At this point, the processor immediately generates a high-level logic interrupt signal, i.e., a feature trigger signal. This signal acts like a camera shutter, instructing the system to stop movement and lock all current data. This method uses an adaptive stopping mechanism based on the rate of change of mechanical forces to replace the traditional fixed displacement or fixed time stopping method, enabling the capture of each subject's unique "elastic limit point" and eliminating undersampling or oversampling problems caused by individual differences.

[0051] S5. In response to the feature trigger signal, terminate the sliding and record the sliding distance from the sliding start point. At the same time, collect data from the sliding start point to the feature trigger point to obtain a complete torque displacement correlation dataset. Optionally, obtaining the complete torque-displacement correlation dataset includes: The pulse count value of the displacement recording unit used to monitor the movement of the evaluation probe is obtained at the moment the feature trigger signal is generated. The algebraic difference between the pulse count value and the initial count value at the sliding starting point is calculated to obtain the sliding distance. The torque sampling sequence and displacement sampling sequence recorded starting from the sliding starting point are retrieved as the original sampling data; The original sampled data is truncated and spatiotemporally aligned based on the displacement endpoint using the sliding distance to generate a complete torque-displacement correlation dataset.

[0052] Specifically, this process is activated the instant the system receives the feature trigger signal. First, the motion control system immediately locks onto the displacement recording unit used to monitor and evaluate probe movement, typically a high-resolution optical or magnetic encoder, and reads the pulse count value at that moment. The system then retrieves the initial count value recorded at the sliding start point from memory and performs an algebraic difference operation on these two count values ​​to obtain the total pulse difference. This pulse difference is multiplied by the encoder's unit pulse equivalent, i.e., the pre-calibrated displacement represented by each pulse, to calculate the precise sliding distance. This sliding distance is a key macroscopic indicator characterizing skin elasticity. This calculation can be expressed as: , in, For the final calculated sliding distance, The pulse count value at the moment the characteristic trigger signal is generated. This is the initial count value at the starting point of the slide. The calibration coefficients for the displacement recording unit are expressed in millimeters per pulse.

[0053] Simultaneously, the data acquisition module responds to the feature trigger signal and stops recording new data. It then retrieves the torque and displacement sampling sequences continuously recorded by the data acquisition buffer, starting from the sliding initiation point. These two sequences are raw time-series data obtained through high-frequency sampling, containing all information from the entire sliding process—that is, the raw sampled data.

[0054] Using the calculated sliding distance as the endpoint benchmark for data processing, interval truncation and spatiotemporal alignment are performed on the original sampled data based on the displacement endpoint. Interval truncation means extracting all data points from the zero time point to the time point corresponding to the feature trigger signal from the two original data sequences, discarding all data after that. Spatiotemporal alignment ensures that in the generated dataset, each torque sample value is precisely paired with its displacement value occurring at the same sampling time, eliminating any possible slight time delay between sensors. Through this processing, a two-dimensional array or similar data structure is finally generated, where each element is a (torque, displacement) data pair. This structured dataset is the complete torque-displacement correlation dataset.

[0055] For example, upon receiving the feature trigger signal, the system reads the current pulse count value of the photoelectric encoder as 1500 and retrieves the initial count value of the sliding start point as 500. The encoder's calibration coefficients are known. mm / pulse, the system calculates the sliding distance Millimeters. Using this 10.0 mm displacement point as the endpoint, the system extracts all raw sampled data from the annular buffer during the time interval from the start of sliding to the generation of the trigger signal. The sampling frequency is 100 Hz, and a total of N sets of data are collected. The system records the torque value at each moment. With the corresponding displacement value By performing one-to-one mapping and spatiotemporal alignment, a coordinate system containing N points is constructed. A sequence of two-dimensional arrays, for example This sequence constitutes the complete torque-displacement correlation dataset required for subsequent analysis. Through rigorous temporal truncation and spatial mapping, this method generates a dataset that eliminates temporal deviations caused by sensor transmission delays, constructing an accurate digital twin model that reflects the dynamic viscoelastic nature of the skin. This provides a clean and complete data foundation for subsequent high-precision feature extraction.

[0056] S6. Based on the sliding distance and the complete torque displacement correlation dataset, feature extraction is performed to obtain skin biomechanical feature parameters; Optionally, the obtained skin biomechanical characteristic parameters include: Numerical integration is performed on the complete torque-displacement correlation dataset to obtain the area under the torque-displacement curve as the first feature parameter. Extract the peak rate of change of the torque change rate before the generation of the feature trigger signal from the complete torque-displacement correlation dataset, and obtain the peak rate of torque change as the second feature parameter; The sliding distance is extracted as a third feature parameter, and combined with the first feature parameter and the second feature parameter to obtain the skin biomechanical feature parameters.

[0057] Specifically, the complete torque-displacement correlation dataset is treated as a function curve composed of discrete points. Using the trapezoidal rule or a similar numerical integration algorithm, the area enclosed by this curve and the displacement axis is calculated. This integral area represents the total amount of externally applied energy absorbed and dissipated by the skin tissue from the initiation of sliding to the characteristic trigger point, physically corresponding to torsional work. This calculation result is defined as the first characteristic parameter. Its calculation can be expressed as: , in, The integral area under the torque-displacement curve. and Let represent the torque value and displacement value of the i-th sampling point in the dataset, respectively. Summation is performed on all data from the second point to the last point.

[0058] The peak rate of change prior to the generation of the feature trigger signal was extracted from the complete torque-displacement correlation dataset. To achieve this, the torque sequence in the dataset was first differencing with respect to displacement to approximate the instantaneous rate of torque change with displacement, i.e., torsional stiffness. Subsequently, the entire rate of change sequence was traversed, and its maximum value was located and recorded. This peak characterizes the moment when the skin's resistance to deformation increases most rapidly during stretching, reflecting the hardness or firmness of the skin tissue. This peak was recorded as the peak rate of torque change and used as the second feature parameter.

[0059] The calculated sliding distance is directly extracted, quantifying the maximum range of skin extension before reaching quasi-static equilibrium. This sliding distance is directly used as the third feature parameter. These three independent values—the integral area as the first feature parameter, the peak torque rate of change as the second feature parameter, and the sliding distance as the third feature parameter—are then vectorized and combined. A mathematical feature vector is constructed according to the dimensionality requirements of the subsequent state evaluation model input layer. .in, Corresponding integral area, Corresponding to the peak rate of torque change. The corresponding sliding distance. This vectorized structure represents the final output of the skin's biomechanical characteristic parameters. It unifies the various physically meaningful measurement indicators into a standard tensor format processed by the adaptive state assessment model, comprehensively and digitally describing the skin's dynamic response in this measurement from three dimensions: energy, stiffness, and extensibility. For example... Figure 3 As shown in the radar chart, the normalized feature vector structure is intuitively displayed by the integral area, the peak torque change rate, and the sliding distance.

[0060] For example, the system performs feature extraction on the acquired complete torque-displacement correlation dataset. First, the first feature parameter is calculated using the trapezoidal rule. In the dataset, two simplified points in a certain segment are (2mm, 0.005Nm) and (3mm, 0.008Nm). The area contribution of this segment is... Joule, summing all segments to get the total area is Microjoules. Secondly, the peak torque change rate was found through differential calculations, and the largest torque increment was discovered within the displacement range of 3mm to 4mm. The slope was then calculated. Newton-meters per meter, record this maximum slope as the second characteristic parameter. Finally, the sliding distance is directly extracted. Meter is used as the third feature parameter. The system combines these three parameters to construct the feature vector to be evaluated. As biomechanical characteristic parameters of the skin, this method abstracts complex continuous waveform data into three physically meaningful dimensions: energy dissipation, stiffness characteristics, and extensibility. This enables the quantitative deconstruction of multidimensional states such as skin laxity, elasticity, and firmness, facilitating algorithmic processing and clinical interpretation.

[0061] S7. Input the skin biomechanical characteristic parameters into the state assessment model for weighted fusion and mapping calculation, and compare them with the classification threshold to obtain the skin state classification label.

[0062] Optionally, the obtained skin condition classification labels include: The skin biomechanical characteristic parameters are input into a state assessment model trained with historical data for processing to obtain feature mapping results. The skin condition index is obtained by weighting and fusing the feature mapping results using the state assessment model. The skin condition index is compared with the classification threshold to obtain the skin condition classification label.

[0063] Specifically, the process begins with the integration of skin biomechanical parameters, i.e., the integral area. Peak torque change rate and sliding distance The resulting feature vectors are fed as input into a lightweight neural network model based on a three-layer feedforward structure. The model's internal structure includes an input layer, a hidden layer with five neurons, and a single-neuron output layer. Its aim is to establish a mapping from the biomechanical feature space to the probability space of skin health status through hierarchical nonlinear transformations. Internally, the input feature vectors are first standardized using Z-scores. This step aims to eliminate dimensional differences; the specific feature mapping calculation formula is as follows: , in, For the standardized first Each feature component The original input features, and These are the mean and standard deviation of the feature in the database statistics, respectively. These two statistics are fixed in the model as preset constants.

[0064] Subsequently, the state evaluation model performs a weighted fusion of the feature mapping results. A computational logic combining fully connected weighted summation and the sigmoid activation function is employed. First, the weighted sum of the hidden layers is calculated, and then the result is passed to the output layer to calculate the final skin state index. The specific nonlinear mapping function is defined as follows: , in, Is the output range within The dimensionless scalar between them These are the standardized features of the input; These are feature weight coefficients. It's a bias term. Here... and The parameters were not randomly assigned, but obtained through an offline supervised learning process. During the model building phase, the backpropagation algorithm was used, with a large amount of skin data labeled with doctors' diagnoses as the training set. Cross-entropy was used as the loss function, and the weights were iteratively updated using gradient descent until the loss function converged, thus obtaining the optimal set of parameters that can best distinguish different skin conditions.

[0065] The calculated skin condition index is compared to a set of preset classification thresholds. These thresholds are set based on the probability distribution characteristics of the Sigmoid function: for example, setting a threshold... and Perform multi-level conditional judgments: If The condition is determined to be "healthy / highly elastic"; if The condition is classified as "sub-healthy / general state"; if The condition is classified as "needs improvement / relaxation". Through this series of comparisons, the descriptive label that best matches the current index range is finally determined, and this label is the final output skin condition classification label.

[0066] For example, the system will use feature vectors The input is fed into a pre-trained state evaluation model. This model is a three-layer neural network: 3 nodes in the input layer, 5 nodes in the hidden layer, and 1 node in the output layer. First, Z-score normalization is performed, and the mean of the training set is used. Standard deviation Then the standardized input , , Next, weighted fusion is performed. The formula for the comprehensive linear weighted sum after hidden layer processing is as follows: Set weights , , bias ,but Finally, the skin condition index is calculated using the Sigmoid function. The index of 0.62 was compared with the classification threshold, setting the "healthy" threshold at 0.75 and the "sub-healthy" range at 0.4 to 0.75. Since 0.4 < 0.62 < 0.75, the system determined and output "sub-healthy / general state" as the final skin condition classification label. Utilizing a non-linear assessment model trained on large datasets, the system can uncover implicit correlations that are difficult to reflect with a single physical indicator. The output standardized classification label not only possesses high objectivity and uniformity but also assists professionals in quickly assessing the degree of skin aging or pathological conditions, thus improving diagnostic efficiency.

[0067] Based on the same inventive concept, such as Figure 4 As shown, the present invention also provides a skin condition assessment system based on torque-triggered data, comprising: The scene parameter matching module is used to obtain information about the skin part to be evaluated, retrieve the scene pattern database to load the corresponding evaluation scene parameter set and skin biomechanical reference model, and obtain scene adaptation parameters. The posture pressure sensing module is used to establish a body surface coordinate system on the skin part to be evaluated according to the scene adaptation parameters, and control the evaluation probe to make initial contact with the skin surface to collect pressure and posture information, and obtain contact pressure data and probe posture data. The dynamic control and regulation module is used to control the movement of the evaluation probe according to the contact pressure data and the probe posture data, and to perform adaptive sliding control and dynamic torque balance regulation during the movement to obtain dynamic torque balance regulation parameters. The feature signal triggering module is used to perform torque curve feature matching based on the dynamic torque balance control parameters and the skin biomechanical reference model to obtain a feature triggering signal. The associated data acquisition module is used to terminate the sliding in response to the feature trigger signal and record the sliding distance from the sliding start point, while collecting data from the sliding start point to the feature trigger point to obtain a complete torque displacement associated dataset. The mechanical feature extraction module is used to extract features based on the sliding distance and the complete torque-displacement correlation dataset to obtain skin biomechanical feature parameters. The state decision classification module is used to input the skin biomechanical feature parameters into the state assessment model for weighted fusion and mapping calculation, and compare them with the classification threshold to obtain the skin state classification label.

[0068] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method can be applied to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0069] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A skin condition assessment method based on torque-triggered data, characterized in that, The method includes: Obtain information about the skin area to be evaluated, retrieve the scene pattern database to load the corresponding evaluation scene parameter set and skin biomechanical reference model, and obtain scene adaptation parameters; Based on the scene adaptation parameters, a body surface coordinate system is established on the skin area to be evaluated, and the evaluation probe is controlled to make initial contact with the skin surface to collect pressure and posture information, thereby obtaining contact pressure data and probe posture data. The evaluation probe is moved according to the contact pressure data and the probe posture data, and adaptive sliding control and dynamic torque balance regulation are performed during the movement to obtain dynamic torque balance regulation parameters. Based on the dynamic torque balance control parameters and the skin biomechanical reference model, torque curve features are matched to obtain feature trigger signals; In response to the feature trigger signal, the sliding is terminated and the sliding distance from the sliding start point is recorded. At the same time, data from the sliding start point to the feature trigger point is collected to obtain a complete torque-displacement correlation dataset. Based on the sliding distance and the complete torque displacement correlation dataset, feature extraction is performed to obtain skin biomechanical feature parameters; The skin biomechanical characteristic parameters are input into the state assessment model for weighted fusion and mapping calculation, and compared with the classification threshold to obtain the skin state classification label.

2. The skin condition assessment method based on torque-triggered data according to claim 1, characterized in that, The obtained scene adaptation parameters include: Based on the skin area information, match and select an evaluation scene mode from the scene mode database; Load the associated initial pressure parameters, expected sliding direction parameters, and skin biomechanical reference model from the evaluation scenario mode, wherein the biomechanical reference model includes an ideal drag torque displacement characteristic curve; By combining the initial pressure parameters, the expected sliding direction parameters, and the skin biomechanical reference model, scene adaptation parameters are obtained.

3. The skin condition assessment method based on torque-triggered data according to claim 1, characterized in that, The obtained contact pressure data and probe attitude data include: The expected sliding direction parameter in the scene adaptation parameters is analyzed, and the direction consistent with the expected sliding direction is defined as the horizontal axis of the body surface coordinate system, and the normal feed axis of the evaluation probe is defined as the vertical axis of the body surface coordinate system to construct the body surface coordinate system. The initial pressure parameter in the scene adaptation parameters is extracted as the preload threshold, and the evaluation probe is controlled to advance along the longitudinal axis of the body surface coordinate system to the skin surface until the real-time value fed back by the pressure sensor reaches the preload threshold and enters the steady state range. The real-time pressure value after entering the steady-state range is read as the contact pressure data, and the tilt angle data of the probe relative to the longitudinal axis of the body surface coordinate system is collected by the attitude sensor set on the evaluation probe as the probe attitude data.

4. The skin condition assessment method based on torque-triggered data according to claim 1, characterized in that, The obtained dynamic torque balance control parameters include: The path following deviation is calculated based on the probe posture data and the virtual reference path established based on the body surface coordinate system. The torque change rate data of the evaluation probe during the sliding process is monitored in real time, compared with the ideal torque change trend defined in the skin biomechanical reference model, and the pressure adjustment amount or direction correction amount is calculated in combination with the path following deviation to obtain the adjusted contact pressure data or the guiding signal used to correct the sliding direction. By combining the torque change rate data, the path following deviation, and the adjusted contact pressure data or the guiding signal, dynamic torque balance control parameters are obtained.

5. The skin condition assessment method based on torque-triggered data according to claim 4, characterized in that, The obtained adjusted contact pressure data or the guiding signal used to correct the sliding direction includes: When the torque change rate data exceeds the upper limit threshold corresponding to the ideal torque change trend, a pressure reduction control command or a prompt to increase path offset is generated. When the torque change rate data is lower than the lower limit threshold corresponding to the ideal torque change trend, a pressure increase control command is generated or a guidance command prompting correction to the virtual reference path is generated.

6. The skin condition assessment method based on torque-triggered data according to claim 1, characterized in that, The obtained feature trigger signal includes: Real-time torque change rate data is extracted from the dynamic torque balance control parameters, and it is determined whether the torque changes into and remains in the range near the zero value used to characterize the stable state, thus obtaining the equilibrium maintenance state. If the duration of the equilibrium state reaches the stability time threshold used to determine the equilibrium state, then the torque curve characteristic triggering condition is satisfied, and a characteristic triggering signal is obtained.

7. The skin condition assessment method based on torque-triggered data according to claim 1, characterized in that, The obtained complete torque-displacement correlation dataset includes: The pulse count value of the displacement recording unit used to monitor the movement of the evaluation probe is obtained at the moment the feature trigger signal is generated. The algebraic difference between the pulse count value and the initial count value at the sliding starting point is calculated to obtain the sliding distance. The torque sampling sequence and displacement sampling sequence recorded starting from the sliding starting point are retrieved as the original sampling data; The original sampled data is truncated and spatiotemporally aligned based on the displacement endpoint using the sliding distance to generate a complete torque-displacement correlation dataset.

8. The skin condition assessment method based on torque-triggered data according to claim 1, characterized in that, The obtained skin biomechanical characteristic parameters include: Numerical integration is performed on the complete torque-displacement correlation dataset to obtain the area under the torque-displacement curve as the first feature parameter. Extract the peak rate of change of the torque change rate before the generation of the feature trigger signal from the complete torque-displacement correlation dataset, and obtain the peak rate of torque change as the second feature parameter; The sliding distance is extracted as a third feature parameter, and combined with the first feature parameter and the second feature parameter to obtain the skin biomechanical feature parameters.

9. The skin condition assessment method based on torque-triggered data according to claim 1, characterized in that, The obtained skin condition classification labels include: The skin biomechanical characteristic parameters are input into a state assessment model trained with historical data for processing to obtain feature mapping results. The skin condition index is obtained by weighting and fusing the feature mapping results using the state assessment model. The skin condition index is compared with the classification threshold to obtain the skin condition classification label.

10. A skin condition assessment system based on torque-triggered data, characterized in that, The system includes: The scene parameter matching module is used to obtain information about the skin part to be evaluated, retrieve the scene pattern database to load the corresponding evaluation scene parameter set and skin biomechanical reference model, and obtain scene adaptation parameters. The posture pressure sensing module is used to establish a body surface coordinate system on the skin part to be evaluated according to the scene adaptation parameters, and control the evaluation probe to make initial contact with the skin surface to collect pressure and posture information, and obtain contact pressure data and probe posture data. The dynamic control and regulation module is used to control the movement of the evaluation probe according to the contact pressure data and the probe posture data, and to perform adaptive sliding control and dynamic torque balance regulation during the movement to obtain dynamic torque balance regulation parameters. The feature signal triggering module is used to perform torque curve feature matching based on the dynamic torque balance control parameters and the skin biomechanical reference model to obtain a feature triggering signal. The associated data acquisition module is used to terminate the sliding in response to the feature trigger signal and record the sliding distance from the sliding start point, while collecting data from the sliding start point to the feature trigger point to obtain a complete torque displacement associated dataset. The mechanical feature extraction module is used to extract features based on the sliding distance and the complete torque-displacement correlation dataset to obtain skin biomechanical feature parameters. The state decision classification module is used to input the skin biomechanical feature parameters into the state assessment model for weighted fusion and mapping calculation, and compare them with the classification threshold to obtain the skin state classification label.

Images