An evaluation system and method for hand rehabilitation detection of burn patients

Abstract

The present invention relates to the technical field of rehabilitation assessment, and discloses an assessment system and method for hand rehabilitation detection of burn patients. Sensors are set to detect the hands of burn patients to identify relevant hand data, and an AR module is set to allow the patient to perform relevant detection actions according to instruction information to obtain a real-time interaction process of AR. The acquired hand data and the data of the real-time interaction process of AR are fused and input into a constructed rehabilitation assessment model to obtain a rehabilitation assessment result, thereby achieving the comprehensiveness and effectiveness of the detection data and greatly improving the accuracy of the rehabilitation assessment result.

Description

Technical Field

[0001] The present invention relates to the technical field of rehabilitation assessment, and in particular to an assessment system and method for hand rehabilitation detection of burn patients. Background Art

[0002] In the field of burns, due to the slowness and complexity of burn recovery, the assessment of the recovery status of burn patients has always been one of the current research hotspots.

[0003] Traditional evaluation methods usually rely on the subjective judgment of doctors and subjective feedback from patients. They lack objectivity and standardization, are easily affected by individual differences and subjective errors, and limit the accurate evaluation of rehabilitation effects.

[0004] At present, sensor technology has gradually been widely used in the field of rehabilitation, providing new ideas and methods for rehabilitation training. For example, sensors such as flexible pressure sensors, electromyography sensors, and inertial measurement units can be used to monitor the patient's hand movements and movements, extract key physiological parameters and movement characteristics, and provide the possibility for objective evaluation of rehabilitation effects. However, current research is mostly focused on the application of a single technology, which still leads to insufficient comprehensiveness and accuracy of rehabilitation evaluation. Summary of the invention

[0005] In view of this, the object of the present invention is to provide an evaluation system and method for hand rehabilitation detection of burn patients, so as to achieve effective evaluation of the hand rehabilitation status of burn patients and improve the accuracy of the evaluation results.

[0006] The first aspect of the present invention discloses an evaluation system for hand rehabilitation detection of burn patients, the evaluation system comprising an indication module, a detection module, a data analysis module, an augmented reality AR module and a display module;

[0007] The indication module is used to issue hand movement indications to the burn patient;

[0008] The detection module is used to set a sensor to detect hand data of a burn patient;

[0009] The augmented reality (AR) module is used to perform AR real-time interaction with the hand of the burn patient during rehabilitation testing based on the hand data and through AR technology, and record the AR real-time interaction process;

[0010] The data analysis module is used to fuse the hand data and the data of the AR real-time interaction process, and analyze and obtain the rehabilitation assessment result of the hand of the burn patient;

[0011] The display module includes displaying the rehabilitation assessment result and the AR real-time interaction process.

[0012] Further, the sensor includes a flexible pressure sensor, an electromyography sensor, 1. An evaluation system for hand rehabilitation detection of burn patients, characterized in that the evaluation system includes an indication module, a detection module, a data analysis module, an augmented reality AR module and a display module;

[0013] The indication module is used to issue hand movement indications to the burn patient;

[0014] The detection module is used to set a sensor to detect hand data of a burn patient;

[0015] The augmented reality (AR) module is used to perform AR real-time interaction with the hand of the burn patient during rehabilitation testing based on the hand data and through AR technology, and record the AR real-time interaction process;

[0016] The data analysis module is used to fuse the hand data and the data of the AR real-time interaction process, and analyze and obtain the rehabilitation assessment result of the hand of the burn patient;

[0017] The display module includes displaying the rehabilitation assessment result and the AR real-time interaction process.

[0018] Inertial measurement unit, surface temperature sensor, flexion sensor, electrode impedance sensor, and 3D motion capture unit.

[0019] Furthermore, the hand data includes hand detection data of the burn patient's hand when it is static and dynamic hand detection data when the hand performs corresponding hand movements according to the instruction module.

[0020] Furthermore, the augmented reality (AR) module includes augmented reality (AR) glasses;

[0021] The augmented reality (AR) module is used to perform AR real-time interaction with the hand of the burn patient during rehabilitation testing based on the hand data and through AR technology, specifically including:

[0022] A virtual hand model of the burn patient is created through an augmented reality (AR) module, and a real-time hand image captured by a three-dimensional motion capture unit is obtained, and the created virtual hand model is superimposed on the captured real-time hand image;

[0023] Acquire dynamic hand detection data of the burn patient in real time, and analyze the dynamic hand detection data to obtain a hand posture analysis result;

[0024] Adjusting the posture of the virtual hand model in real time according to the hand posture analysis result;

[0025] The burn patient observes the real-time posture of the virtual hand model through the AR glasses and adjusts the hand movement according to the indication module.

[0026] Furthermore, the three-dimensional motion capture unit is also used to identify the real object images around the hand of the burn patient;

[0027] The augmented reality (AR) module is also used to merge the real object image with the virtual environment. The burn patient adjusts hand movements according to the indication module and interacts the virtual object with the real object by performing hand movements in the virtual environment seen by the AR glasses.

[0028] Furthermore, the data analysis module also includes a feature extraction submodule, a data fusion submodule and a machine learning submodule;

[0029] The data analysis module is used to fuse the hand data and the data of the AR real-time interaction process, and analyze the rehabilitation assessment results of the burn patient's hand, which specifically include:

[0030] The feature extraction submodule extracts features of hand data and data of the AR real-time interaction process;

[0031] The data fusion submodule fuses the hand data after feature extraction and the data of the AR real-time interaction process through feature fusion technology;

[0032] The fused data is modeled using a machine learning algorithm through the machine learning submodule to obtain a rehabilitation assessment model. Rehabilitation assessment is performed on the hand data acquired in real time and the data of the AR real-time interaction process according to the rehabilitation assessment model to obtain the rehabilitation assessment results of the hands of burn patients.

[0033] Furthermore, the feature fusion technology includes a factor analysis algorithm.

[0034] Furthermore, the machine learning algorithm is a support vector machine algorithm;

[0035] The formula of the prediction function of the rehabilitation assessment model based on the support vector machine algorithm is as follows:

[0036]

[0037] y=f(X)

[0038] Where X is the input feature matrix, which contains the hand data detected by the sensor and the data of the AR real-time interaction process; f(·) is the prediction function of the support vector machine model, and y is the rehabilitation assessment result output by the prediction function; sign(·) is the sign function, which means mapping the output of the function to the category label. If the input value is greater than or equal to zero, the output is a positive category (+1), that is, the rehabilitation index is positive, otherwise the output is a negative category (-1), that is, the rehabilitation index is negative; ∝ iis the Lagrange multiplier of the support vector machine model, which represents the weight of each training sample; i is the label of the training sample, indicating the category of the rehabilitation assessment result; K(x i ,x) is the kernel function, which is used to map the input feature matrix X into a high-dimensional space; i is the feature vector of the training sample; b is the bias term of the model; i represents the i-th sample, and m is the total number of samples.

[0039] The second aspect of the present invention discloses an evaluation method for hand rehabilitation detection of burn patients, the evaluation method is implemented based on the evaluation system for hand rehabilitation detection of burn patients disclosed in the first aspect, and the evaluation method comprises:

[0040] giving instructions on hand movements to burn victims;

[0041] Setting sensors to detect hand data of burn patients;

[0042] Performing AR real-time interaction with the hand of the burn patient during rehabilitation testing through AR technology according to the hand data detected by the sensor, and recording the AR real-time interaction process;

[0043] The hand data and the data of the AR real-time interaction process are fused, and analyzed to obtain a rehabilitation assessment result of the hand of the burn patient;

[0044] The rehabilitation assessment result and the AR real-time interaction process are displayed.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] The present invention detects the hands of burn patients by setting sensors to identify relevant hand data, and sets an AR module to let the patient perform relevant detection actions according to the instruction information to obtain the real-time interaction process of AR, and fuses the acquired hand data and the data of the real-time interaction process of AR, and inputs them into the constructed rehabilitation assessment model to obtain the rehabilitation assessment results, thereby achieving the comprehensiveness and effectiveness of the detection data and greatly improving the accuracy of the rehabilitation assessment results. . BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0048] Figure 1 The present invention is a schematic diagram of the structure of an evaluation system for hand rehabilitation detection of burn patients disclosed in one embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments.

[0050] Embodiment 1

[0051] See also Figure 1 , Figure 1 It is a structural schematic diagram of an evaluation system for hand rehabilitation detection of burn patients disclosed in an embodiment of the present invention, wherein the evaluation system includes an indication module, a detection module, a data analysis module, an augmented reality (AR) module, and a display module;

[0052] The indication module is used to issue hand movement instructions to the burn patient;

[0053] The detection module is used to set sensors to detect hand data of burn patients;

[0054] The augmented reality (AR) module is used to realize AR real-time interaction when performing rehabilitation testing on the hand of a burn patient based on the hand data detected by the above-mentioned sensor and through AR technology, and record the AR real-time interaction process;

[0055] The data analysis module is used to fuse the hand data and the data of the AR real-time interaction process, and analyze and obtain the rehabilitation assessment result of the hand of the burn patient;

[0056] The display module includes displaying the rehabilitation assessment result and the AR real-time interaction process.

[0057] Specifically, in an embodiment of the present invention, it is first necessary to determine the types of rehabilitation status detection actions that need to be performed based on the rehabilitation needs and rehabilitation plan of the burn patient. These types of actions include but are not limited to hand extension, grip strength, flexibility training, etc. Then, based on the determined types of rehabilitation status detection actions, a corresponding rehabilitation status detection action plan is formulated, for example, determining the number of executions, duration, frequency and other parameters of each action, as well as the interval time between actions. Based on the rehabilitation status detection action plan, corresponding action instruction information is generated, and the specific actions that need to be performed are communicated to the burn patient through the instruction module. This can be achieved through display on the display interface in the display module, voice prompts, vibration feedback, etc., to ensure that the patient can clearly receive the instruction information.

[0058] Furthermore, the sensors in the embodiments of the present invention include a flexible pressure sensor, an electromyography sensor, an inertial measurement unit, a surface temperature sensor, a flexion sensor, an electrode impedance sensor, and a three-dimensional motion capture unit.

[0059] Among them, the flexible pressure sensor is used to collect the pressure distribution data of the hand, including the pressure distribution of the fingers and palms, and the pressure data of the hand on the surface of the object, so as to evaluate the contact stability and force transmission between the hand and the object. The electromyography sensor is used to collect the electrical signal data of the hand muscle activity to evaluate the activity degree and coordination of the muscle. The inertial measurement unit is used to collect the acceleration, angular velocity and other motion data of the hand, and is used to analyze the motion trajectory, speed and acceleration and other parameters of the hand. The surface temperature sensor is used to collect the surface temperature data of the hand skin to evaluate the blood circulation and local inflammation. The flexion sensor is used to collect the flexion angle data of the finger joints, which is used to evaluate the range of motion and flexibility of the hand joints. The electrode impedance sensor is used to collect the resistance data between the skin and the electrode, which is used to monitor the edema and skin condition of the hand. The three-dimensional motion capture unit mainly collects the three-dimensional motion trajectory data of the hand through the camera device, including the position, direction and range of motion of the hand. By setting the above sensors on the hands of the burn patient to detect the relevant hand data, the present invention does not specifically limit the setting method of each sensor.

[0060] Furthermore, the present embodiment divides the hand data detected into hand detection data when the hand of the burn patient is static and dynamic hand detection data when the hand performs corresponding hand movements according to the indication module. Specifically, the hand detection data when the hand of the burn patient is static mainly includes the hand detection data when the burn patient remains static before performing relevant movements according to the indication module, such as the surface temperature data of the hand skin, the resistance data between the skin and the electrode, etc. When the patient performs relevant hand movements according to the indication module, the surface temperature data of the hand skin, the resistance data between the skin and the electrode, and other data are also detected to be compared with the static hand detection data, so as to improve the accuracy of the final evaluation results of the patient's hand rehabilitation.

[0061] Furthermore, the augmented reality AR module includes augmented reality AR glasses. The augmented reality AR module is used to realize AR real-time interaction when performing rehabilitation testing on the hands of burn patients based on hand data and through AR technology, specifically including:

[0062] A virtual hand model of the burn patient is created through an augmented reality (AR) module, and a real-time image of the hand captured by a three-dimensional motion capture unit is obtained, and the created virtual hand model is superimposed on the captured real-time image of the hand.

[0063] The dynamic hand detection data of the burn patient is obtained in real time, and the hand posture analysis results are obtained based on the dynamic hand detection data;

[0064] Adjust the posture of the virtual hand model in real time according to the hand posture analysis results;

[0065] Burn patients observe the real-time posture of the virtual hand model through AR glasses and adjust their hand movements according to the instruction module.

[0066] Specifically, in the embodiment of the present invention, the burn patient can obtain the instruction information issued by the instruction module by wearing AR glasses, and complete the relevant actions according to the instruction information. At the same time, the patient can also synchronously watch the completion of his own actions through the AR glasses.

[0067] Furthermore, the 3D motion capture unit is also used to identify the real object images around the burn patient's hands. The augmented reality AR module is also used to merge the real object images with the virtual environment. The burn patient adjusts the hand movements according to the instruction module and interacts the virtual objects with the real objects by performing hand movements in the virtual environment seen by the AR glasses.

[0068] In this embodiment, the image of the real object is integrated with the virtual environment through the augmented reality AR module. The patient can experience different interesting scenes while wearing AR glasses and making hand movements according to the instructions. For example, when the patient receives the action instruction of pinching and moving the hand, the virtual sphere can be pinched and moved in the virtual environment to the real desktop integrated with the virtual environment. Through the above settings, when the patient is undergoing rehabilitation detection, he can accurately perform the rehabilitation detection actions and experience the fun of the detection process. In this case, the patient can also effectively interact during the rehabilitation detection and evaluation, improve the comprehensiveness of the detection data, and further improve the accuracy of the rehabilitation evaluation results.

[0069] Furthermore, the data analysis module also includes a feature extraction submodule, a data fusion submodule and a machine learning submodule.

[0070] The data analysis module is used to fuse the hand data and the data of the AR real-time interaction process, and analyze the rehabilitation assessment results of the burn patient's hand, which specifically include:

[0071] The feature extraction submodule is used to extract features from hand data and data of the AR real-time interaction process;

[0072] The data fusion submodule fuses the hand data after feature extraction operation and the data of AR real-time interaction process through feature fusion technology.

[0073] The fused data is modeled using a machine learning algorithm through the machine learning submodule to obtain a rehabilitation assessment model. Rehabilitation assessment is performed on the hand data acquired in real time and the data of the AR real-time interaction process according to the rehabilitation assessment model to obtain the rehabilitation assessment results of the hands of burn patients.

[0074] Furthermore, the feature fusion technique includes a factor analysis algorithm.

[0075] Specifically, in the feature extraction stage, we extract features related to rehabilitation assessment from the hand data detected by the sensor and the data of the AR real-time interaction process, such as the range of motion of the hand, force distribution, posture changes, etc., and organize the extracted features into a feature matrix as the original feature matrix, where each row represents a sample, such as a data record of the hand of a burn patient, and each column represents a feature. After completing the data preparation, the original feature matrix is ​​modeled using the factor analysis model to obtain the factor loading matrix, factor score matrix, and error matrix, and the factor loading matrix and factor score matrix are estimated by minimizing the error. Among them, the factor loading matrix represents the relationship between each feature and the latent factor, and the factor score matrix represents the score of each sample on each latent factor.

[0076] After factor analysis, the factor scores of each sample on each potential factor are obtained, and these scores are used as new features together with the original hand data detected from the sensor and the data of the AR real-time interaction process for data fusion. For example, the original feature matrix contains four features: pressure, temperature, electromyographic signals, and the patient's AR interaction response, while the factor score matrix contains two factors: motion factor and pain factor. At this time, the original feature matrix and the factor score matrix are concatenated to form a new feature matrix, in which the new features are pressure, temperature, electromyographic signals, the patient's AR interaction response, and the two factor scores.

[0077] Furthermore, the machine learning algorithm is a support vector machine algorithm.

[0078] The support vector machine algorithm works by finding an optimal hyperplane in the data set for classification or regression, and achieves this goal by maximizing the interval between the hyperplane and samples of different categories while minimizing the number of misclassified samples.

[0079] In this embodiment, in the scenario of rehabilitation assessment, the hand data detected by the sensor and the data of the AR real-time interaction process are used as input features, and the rehabilitation assessment results are used as labels, and SVM is used to establish a classification model. The goal of the model is to predict the rehabilitation assessment results based on the input data, such as judging whether the hand rehabilitation is good or whether further intervention is needed.

[0080] Specifically, the formula of the prediction function of the rehabilitation assessment model based on the support vector machine algorithm is as follows:

[0081]

[0082] y=f(X)

[0083] Where X is the input feature matrix, which contains the hand data detected by the sensor and the data of the AR real-time interaction process; f(·) is the prediction function of the support vector machine model, and y is the rehabilitation assessment result output by the prediction function; sign(·) is the sign function, which means mapping the output of the function to the category label. If the input value is greater than or equal to zero, the output is a positive category (+1), that is, the rehabilitation index is positive, otherwise the output is a negative category (-1), that is, the rehabilitation index is negative; ∝ i is the Lagrange multiplier of the support vector machine model, which represents the weight of each training sample; i is the label of the training sample, indicating the category of the rehabilitation assessment result; K(x i ,x) is the kernel function, which is used to map the input feature matrix X into a high-dimensional space; i is the feature vector of the training sample; b is the bias term of the model; i represents the i-th sample, and m is the total number of samples.

[0084] SVM is suitable for high-dimensional data, and in this embodiment, the hand data detected by the sensor and the data recorded by AR contain multiple features, which constitute a high-dimensional feature space, and SVM can be effectively processed. In addition, in this embodiment, the original features are mapped to the high-dimensional space by using the kernel function, so as to achieve the processing of nonlinear problems. In rehabilitation assessment, there may be complex nonlinear relationships, and SVM can capture such relationships through the kernel function to improve the accuracy of the model.

[0085] When rehabilitation outcome assessment is required, the fused features are input into the model, and the rehabilitation assessment result is obtained using the model's prediction function. This result may be a continuous value of a positive category (+1) or a negative category (-1), or a discrete value of a positive category (+1) or a negative category (-1).

[0086] For example, when the result predicted by the rehabilitation assessment model is a continuous value, we can interpret it as the degree of rehabilitation progress, which means that the result output by the model can represent the specific state or degree of progress of the hand rehabilitation of the burn patient, such as a percentage value or a continuous score. Assume that the result output by the rehabilitation assessment model is a value between -100 and 100, indicating the degree of hand rehabilitation of the burn patient, where -100 means no rehabilitation at all and 100 means complete rehabilitation. Medical staff can understand the rehabilitation progress of the burn patient based on this result, so as to adjust the rehabilitation plan or provide personalized treatment recommendations. If the rehabilitation assessment result output by the model is 80, medical staff can interpret it as the hand rehabilitation of the burn patient is good, but there is still a certain degree of rehabilitation space, and rehabilitation training and treatment can be continued to further improve the rehabilitation level.

[0087] On the other hand, when the result predicted by the rehabilitation assessment model is a discrete value, we can interpret it as a classification of rehabilitation results. This means that the result output by the model indicates which category the rehabilitation status of the burn patient belongs to, such as "good", "general", "intervention required", etc. Suppose the output results of the rehabilitation assessment model are three discrete categories: "good", "general" and "intervention required". Medical staff can judge the rehabilitation status of burn patients based on the prediction results of the model and take corresponding measures. If the result output by the model is "intervention required", it means that the rehabilitation of the burn patient is poor and needs to strengthen rehabilitation training or provide more medical intervention.

[0088] Finally, the display module is used to display the rehabilitation assessment results to medical staff or patients. At the same time, these results are interpreted to help medical staff develop personalized rehabilitation plans or provide suggestions.

[0089] Embodiment 2

[0090] The second aspect of the present invention discloses an evaluation method for hand rehabilitation detection of burn patients. The evaluation method is implemented based on the evaluation system for hand rehabilitation detection of burn patients disclosed in the first aspect, and the evaluation method comprises:

[0091] giving instructions on hand movements to burn victims;

[0092] Setting sensors to detect hand data of burn patients;

[0093] According to the hand data detected by the sensor, AR technology is used to interact with the burn patient's hand in real time during rehabilitation testing, and the AR real-time interaction process is recorded;

[0094] The hand data and the data of the AR real-time interaction process are fused and analyzed to obtain the rehabilitation assessment results of the burn patient's hand;

[0095] Display rehabilitation assessment results and AR real-time interaction process.

[0096] Furthermore, the sensor includes a flexible pressure sensor, an electromyography sensor, an inertial measurement unit, a surface temperature sensor, a flexion sensor, an electrode impedance sensor and a three-dimensional motion capture unit.

[0097] Furthermore, the hand data includes hand detection data of the burn patient's hand when it is static and dynamic hand detection data when the hand performs corresponding hand movements according to the instruction module.

[0098] Furthermore, AR real-time interaction is achieved when conducting rehabilitation tests on the hands of burn patients based on hand data through AR technology, including:

[0099] A virtual hand model of the burn patient is created through an augmented reality (AR) module, and a real-time hand image captured by a three-dimensional motion capture unit is obtained, and the created virtual hand model is superimposed on the captured real-time hand image;

[0100] The dynamic hand detection data of the burn patient is obtained in real time, and the hand posture analysis results are obtained based on the dynamic hand detection data;

[0101] Adjust the posture of the virtual hand model in real time according to the hand posture analysis results;

[0102] Burn patients observe the real-time posture of the virtual hand model through AR glasses and adjust their hand movements according to the instructions.

[0103] Furthermore, the real object images around the hand of the burn patient are also recognized by a three-dimensional motion capture unit;

[0104] The real object image is integrated with the virtual environment through the augmented reality (AR) module, and the burn patient adjusts hand movements according to the instruction information and interacts the virtual object with the real object by performing hand movements in the virtual environment seen by the AR glasses.

[0105] Furthermore, the machine learning algorithm is a support vector machine algorithm;

[0106] The formula of the prediction function of the rehabilitation assessment model based on the support vector machine algorithm is as follows:

[0107]

[0108] y=f(X)

[0109] Where X is the input feature matrix, which contains the hand data detected by the sensor and the data of the AR real-time interaction process; f(·) is the prediction function of the support vector machine model, and y is the rehabilitation assessment result output by the prediction function; sign(·) is the sign function, which means mapping the output of the function to the category label. If the input value is greater than or equal to zero, the output is a positive category (+1), that is, the rehabilitation index is positive, otherwise the output is a negative category (-1), that is, the rehabilitation index is negative; ∝ i is the Lagrange multiplier of the support vector machine model, which represents the weight of each training sample; i is the label of the training sample, indicating the category of the rehabilitation assessment result; K(x i ,x) is the kernel function, which is used to map the input feature matrix X into a high-dimensional space; i is the feature vector of the training sample; b is the bias term of the model; i represents the i-th sample, and m is the total number of samples.

[0110] It should be noted that the specific implementation process of the second embodiment is similar to that of the first embodiment and will not be repeated in the second embodiment.

[0111] Finally, it should be noted that the evaluation system and method for hand rehabilitation detection of burn patients disclosed in the embodiment of the present invention only disclose the preferred embodiments of the present invention, which are only used to illustrate the technical scheme of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical schemes described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical schemes from the spirit and scope of the technical schemes of the embodiments of the present invention.

Claims

1. An evaluation system for hand rehabilitation detection of burn patients, characterized in that: The evaluation system includes an indication module, a detection module, a data analysis module, an augmented reality (AR) module, and a display module; The indication module is used to issue hand movement indications to the burn patient; The detection module is used to set a sensor to detect hand data of a burn patient; The augmented reality (AR) module is used to realize AR real-time interaction when performing rehabilitation testing on the hand of the burn patient according to the hand data and through AR technology, and record the AR real-time interaction process; The data analysis module is used to fuse the hand data and the data of the AR real-time interaction process, and analyze and obtain the rehabilitation assessment result of the hand of the burn patient; The display module includes displaying the rehabilitation assessment result and the AR real-time interaction process; The hand data includes hand detection data of the burn patient's hand when it is static and dynamic hand detection data when the hand performs corresponding hand movements according to the indication module; The sensors include a flexible pressure sensor, an electromyography sensor, an inertial measurement unit, a surface temperature sensor, a flexion sensor, an electrode impedance sensor, and a three-dimensional motion capture unit; The augmented reality (AR) module includes augmented reality (AR) glasses; The augmented reality (AR) module is used to realize AR real-time interaction when performing rehabilitation testing on the hand of a burn patient according to the hand data and through AR technology, and specifically includes: A virtual hand model of the burn patient is created through an augmented reality (AR) module, and a real-time hand image captured by a three-dimensional motion capture unit is obtained, and the created virtual hand model is superimposed on the captured real-time hand image; Acquire dynamic hand detection data of the burn patient in real time, and analyze the dynamic hand detection data to obtain a hand posture analysis result; Adjusting the posture of the virtual hand model in real time according to the hand posture analysis result; The burn patient observes the real-time posture of the virtual hand model through AR glasses and adjusts the hand movement according to the instruction module; The three-dimensional motion capture unit is also used to identify the real object images around the hand of the burn patient; The augmented reality (AR) module is further used to merge the real object image with the virtual environment, and the burn patient adjusts the hand movements according to the indication module and interacts the virtual object with the real object by performing hand movements in the virtual environment seen by the AR glasses; The data analysis module also includes a feature extraction submodule, a data fusion submodule, and a machine learning submodule; The data analysis module is used to fuse the hand data and the data of the AR real-time interaction process, and analyze the rehabilitation assessment results of the burn patient's hand, which specifically include: The feature extraction submodule performs feature extraction on the hand data and the data of the AR real-time interaction process to obtain an original feature matrix, in which each row represents a sample and each column represents a feature; The data fusion submodule is used to fuse the hand data after feature extraction and the data of the AR real-time interaction process; The machine learning submodule uses a machine learning algorithm to model the fused data to obtain a rehabilitation assessment model, and performs rehabilitation assessment on the hand data obtained in real time and the data of the AR real-time interaction process according to the rehabilitation assessment model to obtain a rehabilitation assessment result of the hand of the burn patient; The feature fusion technology includes a factor analysis algorithm, which uses a factor analysis model to model the original feature matrix to obtain a factor loading matrix, a factor score matrix and an error matrix, and estimates the factor loading matrix and the factor score matrix by minimizing the error; wherein the factor loading matrix represents the relationship between each feature and the latent factor, and the factor score matrix represents the score of each sample on each latent factor; After performing factor analysis, the original feature matrix and factor score matrix are concatenated to form a new feature matrix.

2. The evaluation system for hand rehabilitation detection of burn patients according to claim 1, characterized in that: The machine learning algorithm is a support vector machine algorithm; The formula of the prediction function of the rehabilitation assessment model based on the support vector machine algorithm is as follows: y=f(X) Where X is the input feature matrix, which contains the hand data detected by the sensor and the data of the AR real-time interaction process; f(·) is the prediction function of the support vector machine model, and y is the rehabilitation assessment result output by the prediction function; sign(·) is the sign function, which means mapping the output of the function to the category label. If the input value is greater than or equal to zero, the output is a positive category (+1), that is, the rehabilitation index is positive, otherwise the output is a negative category (-1), that is, the rehabilitation index is negative; ∝ i is the Lagrange multiplier of the support vector machine model, which represents the weight of each training sample; i is the label of the training sample, indicating the category of the rehabilitation assessment result; K(x i ,x) is the kernel function, which is used to map the input feature matrix X into a high-dimensional space; i is the feature vector of the training sample; b is the bias term of the model; i represents the i-th sample, and m is the total number of samples.

Images