Wearable rehabilitation treatment equipment and method for generating treatment parameters for wearable rehabilitation treatment equipment

Abstract

The embodiment of the invention aims to provide a method for generating treatment parameters for wearable rehabilitation treatment equipment through a neural network model and the wearable rehabilitation treatment equipment. Wherein the neural network model is trained according to a sample data set, and each piece of sample data in the sample data set is labeled with a first diagnosis parameter and a corresponding first treatment parameter; according to a second diagnosis parameter of the user, a corresponding second treatment parameter is generated through the neural network model, and the second treatment parameter is used for the wearable rehabilitation treatment equipment; and updating the second diagnosis parameter according to treatment feedback data of the wearable rehabilitation treatment equipment used by the user. According to the wearable rehabilitation treatment equipment, through the innovative application of combining the user diagnosis parameters and the neural network training model, the treatment parameters are dynamically adjusted, and the wearable rehabilitation treatment equipment is remarkably different from a traditional product.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the fields of wearable devices and rehabilitation medicine, and more particularly to a technology for generating treatment parameters for a wearable rehabilitation therapy device through a neural network model. Background Art

[0002] Low back pain is one of the most common chronic diseases worldwide, with about 80% of people experiencing it in their lifetime. Global medical and economic losses related to low back pain are estimated to exceed hundreds of billions of dollars. There is a huge market demand for products to treat low back pain, but existing treatments are usually lacking in personalization and precision. In addition, current low back pain treatments mainly rely on medication, physical therapy or surgical intervention, which have problems such as unstable efficacy, long treatment cycles, high costs or severe side effects.

[0003] In the process of using low back pain treatment products, many low back pain patients also face similar dilemmas: on the one hand, they lack sufficient medical knowledge to choose appropriate treatment products; on the other hand, the functional design of these products often lacks scientific basis and cannot meet personalized treatment needs.

[0004] When faced with many product choices, due to the lack of professional guidance, users are very blind in the purchase process and find it difficult to judge which product is suitable for their own situation. There is no scientific guidance for the purchase of home physiotherapy equipment, which is more of a marketing, advertising promotion, and ranking according to the comprehensive sales volume on the Internet. Consumers use it blindly after purchase, and only adjust it according to their own feelings, rather than using it scientifically according to their own physical conditions. The use of home physiotherapy equipment is convenient but not scientific enough, and it is more of a blind operation by users. Moreover, during the use process, it is all stand-alone operation, or the use parameters of the equipment are debugged using a mobile phone, and no intelligent debugging is performed based on the results of the patient's physical assessment. For example, in actual use, the purchased therapeutic belt only provides fixed preset modes, and cannot be adjusted according to the specific type of low back pain, the degree of pain, or personal needs, and the effect is greatly reduced.

[0005] The traditional treatment methods for low back pain in the industry mainly rely on the idea of ​​"fixed mode + passive treatment", ignoring the importance of individualized needs and dynamic feedback of patients. These pain points not only lead to serious product homogeneity in the market, but also significantly limit the treatment effect and user experience. Common treatment methods for low back pain are as follows:

[0006] 1. Fixed-mode therapeutic belts: This type of product uses a preset fixed treatment mode, usually including simple heat compress, vibration, low-frequency electrical stimulation and other functions. These modes were designed without full consideration of individual differences among patients, and can only provide standardized treatments. There is a lack of targeted adjustments for different types of low back pain (such as acute, chronic, neurogenic low back pain, etc.) or stages of the disease, resulting in limited treatment effects.

[0007] 2. Physical therapy equipment: Although large-scale clinical physical therapy equipment (such as medium frequency interference electrotherapy equipment, pulse electrotherapy equipment, etc.) can provide relatively professional treatment, the equipment is expensive and complicated to operate, and requires professional guidance. Such equipment is usually limited to hospitals or rehabilitation institutions and cannot meet the needs of patients for home treatment.

[0008] 3. Dependence on surgery and drug treatment: For patients with low back pain, especially those with severe low back pain, surgical intervention and long-term drug treatment are common options. However, surgery is costly, risky, and has the possibility of postoperative complications; while drug treatment may cause addiction or side effects. These methods not only cause pain to patients, but also increase the burden on the medical system.

[0009] These low back pain treatment methods have the following pain points:

[0010] 1. Lack of personalization: Most devices are unable to adjust treatment parameters based on the patient’s diagnostic data, ignoring individual differences in low back pain patients in terms of etiology, course of disease, pain threshold, etc. Fixed-mode, one-size-fits-all treatment methods have resulted in many patients experiencing poor results or even giving up on using the device.

[0011] 2. Blind purchase and use: There are many types of low back pain treatment equipment on the market, but there is a lack of unified industry standards. Consumers often make purchases based on the comprehensive online sales rankings, without obtaining scientific guidance, and the product instructions lack detailed guidance for different symptoms, making it difficult for patients to trust or operate correctly when using them.

[0012] 3. Lack of data support and feedback mechanism: Existing products usually only provide one-way treatment functions, lack real-time monitoring and data analysis capabilities, and cannot achieve dynamic adjustment or optimization. Patients cannot get clear feedback during use and do not know whether the treatment effect meets expectations.

[0013] 4. Uneven distribution of medical resources: Some patients cannot receive effective treatment at the grassroots level or at home due to lack of professional medical conditions, and the effectiveness of existing home treatment equipment is limited, making it difficult to meet patients' needs for high-quality treatment. Summary of the invention

[0014] The purpose of each embodiment of the present disclosure is to provide a method for generating treatment parameters for a wearable rehabilitation therapy device through a neural network model and a wearable rehabilitation therapy device.

[0015] According to one aspect of the present disclosure, there is provided

[0016] A method for generating treatment parameters for a wearable rehabilitation treatment device through a neural network model, wherein the method comprises the following steps:

[0017] Training a neural network model according to a sample data set, wherein each sample data in the sample data set is labeled with a first diagnostic parameter and a corresponding first treatment parameter;

[0018] Obtaining a second diagnostic parameter of the user;

[0019] Inputting the second diagnostic parameter into the trained neural network model to obtain a corresponding second treatment parameter;

[0020] transmitting the second treatment parameter to a corresponding treatment module of the wearable rehabilitation treatment device to perform rehabilitation treatment on the user;

[0021] The method further comprises the following steps:

[0022] According to the treatment feedback data of the user based on the second treatment parameter, the second treatment parameter is updated for use in the next rehabilitation treatment of the user.

[0023] According to one aspect of the present disclosure, there is also provided

[0024] A wearable rehabilitation therapy device includes a control module, a communication module, a neural network model and one or more treatment modules; wherein:

[0025] The communication module is used to receive a second diagnostic parameter of the user and report treatment feedback data of the user based on the second treatment parameter;

[0026] The neural network model is used to obtain a corresponding second treatment parameter according to the second diagnostic parameter, wherein the second diagnostic parameter is also updated according to treatment feedback data of the user using the wearable rehabilitation therapy device; wherein the neural network model is trained according to a sample data set, each sample data in the sample data set is annotated with a first diagnostic parameter and a corresponding first treatment parameter; and the neural network model is optimized according to the updated sample data set, wherein the sample data set is updated according to the treatment feedback data of one or more users using the wearable rehabilitation therapy device;

[0027] The control module is used to transmit the second treatment parameter to the corresponding treatment module, so that each treatment module works according to its corresponding second treatment parameter.

[0028] According to one aspect of the present disclosure, a method for generating treatment parameters for a wearable rehabilitation treatment device through a neural network model is also provided.

[0029] The method comprises the following steps:

[0030] Training the neural network model according to a sample data set, wherein each sample data in the sample data set is annotated with a first diagnostic parameter and a corresponding first treatment parameter;

[0031] According to the second diagnostic parameter of the user, generating corresponding second treatment parameters through the neural network model for use in the wearable rehabilitation treatment device;

[0032] updating the second diagnostic parameter according to the treatment feedback data of the user using the wearable rehabilitation therapy device, so as to be used by the neural network model to generate the corresponding second treatment parameter for the user next time;

[0033] The method further comprises the following steps:

[0034] The neural network model is optimized according to the updated sample data set, and the sample data set is updated according to treatment feedback data of one or more users using the wearable rehabilitation therapy device.

[0035] The wearable rehabilitation therapy device disclosed in the present invention can output personalized treatment plans through the neural network model intelligently according to the user's specific diagnostic parameters and disease classification, including the treatment mode, waveform, frequency, hot compress function, etc. of medium and low frequency electrotherapy, significantly improving the treatment effect while reducing risks. This technological advancement can not only improve the user experience, but also reduce the burden on medical resources.

[0036] The wearable rehabilitation therapy device disclosed in the present invention can be widely used in scenarios such as home rehabilitation therapy, sports injury repair, and long-term low back pain management. Its portability and intelligent features make it particularly suitable for daily use, and can provide patients with accurate and efficient non-invasive treatment, while significantly reducing the medical costs of patients.

[0037] The wearable rehabilitation therapy device disclosed in the present invention combines user diagnostic parameters with the innovative application of a neural network training model to dynamically adjust treatment parameters to adapt to different types of low back pain and the patient's recovery process, while preventing users from blindly selecting parameters and reducing the risk of misoperation or over-reliance on the user's own judgment. The wearable rehabilitation therapy device disclosed in the present invention can not only improve the treatment effect, but also continuously optimize the model performance through data feedback, forming a technical barrier and becoming a leader in the high-end rehabilitation equipment market.

[0038] The wearable rehabilitation device disclosed in this paper breaks through the technical bottleneck of traditional pain treatment equipment by introducing a neural network model, and provides an efficient and accurate solution for personalized treatment. This not only brings good news to patients, but also injects innovative power into the rehabilitation equipment industry, and has extremely high market value and social significance.

[0039] By combining the patient's diagnostic parameters and data and using neural network technology to train the model, wearable rehabilitation therapy equipment can dynamically generate personalized treatment plans based on the patient's specific disease characteristics. This will not only allow patients to get rid of the dilemma of "blindly selecting" and "blindly using" rehabilitation therapy products on the market, but also greatly improve the scientificity and effectiveness of treatment.

[0040] Compared with the traditional fixed-mode treatment belt, the wearable rehabilitation therapy device disclosed in the present invention can accurately match the patient's condition and significantly improve the treatment effect. In addition, the neural network model used in the wearable rehabilitation therapy device disclosed in the present invention has the ability to learn and optimize, and can continuously improve the treatment plan as the patient's usage data accumulates, providing scientific guidance. The wearable rehabilitation therapy device disclosed in the present invention breaks the limitation of traditional products that "one prescription for one thousand people" and provides a new intelligent and personalized solution for the rehabilitation treatment of specific parts, meeting the patient's needs for efficient and convenient treatment, and at the same time has broad market application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Other features, objects and advantages of the present disclosure will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0042] Figure 1 A flow chart of a method for training a neural network model for generating treatment parameters for a user according to an embodiment of the present disclosure is shown.

[0043] The same or similar reference numerals in the drawings represent the same or similar components. DETAILED DESCRIPTION

[0044] The specific embodiments of the present disclosure will be further described below in conjunction with the accompanying drawings.

[0045] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments of the present disclosure are described as devices represented by block diagrams and processes or methods represented by flow charts. Although the flow charts describe the operating processes of the various embodiments of the present disclosure as sequential processing, many of the operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the operations can be rearranged. The processes of the various embodiments of the present disclosure can be terminated when their operations are completed, but can also include additional steps not shown in the flow charts. The processes of the various embodiments of the present disclosure can correspond to methods, functions, procedures, subroutines, subprograms, etc.

[0046] The methods illustrated by the flow charts and the devices illustrated by the block diagrams discussed below may be implemented by hardware, software, firmware, middleware, microcode, hardware description languages, or any combination thereof. When implemented in software, firmware, middleware, or microcode, program code or code segments that perform the necessary tasks may be stored in a machine or computer readable medium such as a storage medium. (One or more) processors may perform the necessary tasks.

[0047] Similarly, it will also be understood that any flow charts, flow diagrams, state transition diagrams, and the like represent various processes that can be fully described as program code stored in a computer readable medium and thereby executed by a computer device or processor, whether or not such computer device or processor is explicitly shown.

[0048] As used herein, the term "storage medium" may refer to one or more devices for storing data, including read-only memory (ROM), random access memory (RAM), magnetic RAM, kernel memory, magnetic disk storage media, optical storage media, flash memory devices and / or other machine-readable media for storing information. The term "computer-readable medium" may include, but is not limited to, portable or fixed storage devices, optical storage devices, and various other media capable of storing and / or containing instructions and / or data.

[0049] A code segment may represent a procedure, function, subprogram, program, routine, subroutine, module, software package, class, or any combination of instructions, data structures, or program descriptions. A code segment may be coupled to another code segment or hardware circuit by passing and / or receiving information, data, arguments, parameters, or storage contents. Information, arguments, parameters, data, etc. may be passed, forwarded, or transmitted via any suitable means including storage sharing, information passing, token passing, network transmission, etc.

[0050] The "computer device" referred to in this context refers to an electronic device that can perform predetermined processing procedures such as numerical calculations and / or logical calculations by running predetermined programs or instructions, and may include at least a processor and a memory, wherein the processor executes program instructions pre-stored in the memory to perform the predetermined processing process, or the predetermined processing process is performed by hardware such as ASIC, FPGA, DSP, or a combination of the above two.

[0051] The above-mentioned "computer device" is usually expressed in the form of a general-purpose computer device, and its components may include but are not limited to: one or more processors or processing units, system memory. The system memory may include a computer-readable medium in the form of a volatile memory, such as a random access memory (RAM) and / or a cache memory. The "computer device" may further include other removable / non-removable, volatile / non-volatile computer-readable storage media. The memory may include at least one computer program product having a set (e.g., at least one) of program modules that are configured to perform the functions and / or methods of the various embodiments of the present disclosure. The processor executes various functional applications and data processing by running the program stored in the memory.

[0052] For example, a computer program for executing various functions and processes of multiple embodiments of the present disclosure is stored in the memory, and when the processor executes the corresponding computer program, multiple embodiments of the present disclosure are implemented.

[0053] Typically, the computer device may be, for example, a user device or a network device, or even a combination of the two. The user device includes, but is not limited to, a personal computer (PC), a laptop, a mobile terminal, etc. The mobile terminal includes, but is not limited to, a smart phone, a tablet computer, etc. The network device includes, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud consisting of a large number of computers or network servers based on cloud computing (Cloud Computing), wherein cloud computing is a type of distributed computing, a super virtual computer consisting of a group of loosely coupled computer sets. The computer device may be operated alone to implement the various embodiments of the present disclosure, or may be connected to a network and implement the various embodiments of the present disclosure through interactive operations with other computer devices in the network. The network in which the computer device is located includes, but is not limited to, the Internet, a wide area network, a metropolitan area network, a local area network, a VPN network, etc.

[0054] It should be noted that the user equipment, network equipment and network are only examples. Other existing or future computing devices or networks that are applicable to the embodiments of the present disclosure should also be included in the scope of protection of the present disclosure and are included here by reference.

[0055] The specific structural and functional details disclosed herein are merely representative and are for the purpose of describing exemplary embodiments of the present disclosure. However, the various embodiments of the present disclosure may be implemented in many alternative forms and should not be interpreted as being limited to only the embodiments set forth herein.

[0056] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0057] The terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments. Unless the context clearly indicates otherwise, the singular forms "one", "one" and "item" used herein are also intended to include plural numbers. It should also be understood that the terms "include" and / or "comprise" used herein specify the existence of stated features, integers, steps, operations, units and / or components, without excluding the existence or addition of one or more other features, integers, steps, operations, units, components and / or combinations thereof.

[0058] It should also be noted that, in some alternative implementations, the functions / actions mentioned may occur in a different order than that indicated in the accompanying drawings. For example, depending on the functions / actions involved, two figures shown in succession may actually be performed substantially simultaneously or may sometimes be performed in the reverse order.

[0059] Multiple embodiments of the present disclosure provide a wearable rehabilitation therapy device. Here, the wearable rehabilitation therapy device is a portable medical device that can perform precise physical therapy on specific parts (such as waist, neck, etc.) through a variety of treatment modules (such as electrical stimulation module, hot compress module, vibration massage module, etc.). The device can generate a personalized treatment plan based on the user's diagnostic parameters (such as pain type, intensity, motor function, muscle strength, core endurance time, etc.), which is suitable for rehabilitation treatment of different parts such as waist, neck, knee joints, etc., supports modular design, and is convenient for adaptation to different application scenarios.

[0060] In addition to the treatment module, the wearable rehabilitation therapy device may also be configured with a communication module to communicate with a user device. The user device, for example, may be a mobile phone or a smart watch, in which an APP or applet (hereinafter referred to as a "rehabilitation therapy applet") for data transmission with the wearable rehabilitation therapy device is installed.

[0061] According to some embodiments, the rehabilitation therapy applet obtains the user's diagnostic parameters and passes them to the communication module of the wearable rehabilitation therapy device. The wearable rehabilitation therapy device determines the corresponding treatment parameters through a neural network model based on the diagnostic parameters, and passes the treatment parameters to each corresponding treatment module through the control module to perform rehabilitation treatment on specific parts of the user.

[0062] According to other embodiments, the neural network model can also be integrated into the rehabilitation therapy applet, or arranged in the cloud and implemented through cloud computing. According to the user's diagnostic parameters, the neural network model determines the corresponding treatment parameters. The treatment parameters can be directly transmitted to the wearable rehabilitation therapy device, or transmitted from the cloud to the wearable rehabilitation therapy device via the user device. The wearable rehabilitation therapy device transmits the treatment parameters to each corresponding treatment module through the control module.

[0063] In addition, as a wearable device, the wearable rehabilitation therapy device is a lightweight and wearable smart device as a whole, such as a smart belt, which adopts ergonomic design and flexible materials that fit the body curve of specific parts, such as the waist curve, to ensure comfortable wearing. Extension belts can also be configured to suit users of different body shapes. The appearance and functional design of the device are suitable for daily wear. It is light in weight, small in size, and easy to carry. It can be used at any time at home, in the office, or on the go. It is a health management device that combines medical and life attributes.

[0064] The wearable rehabilitation device is connected to the rehabilitation applet via Bluetooth. Users can view diagnostic results, manually adjust treatment modes, or monitor usage data through the applet. The device supports USB charging and can last for multiple days on a single charge.

[0065] Wearable rehabilitation therapy devices are equipped with various treatment modules, such as electrical stimulation units, graphene hot compress devices, and vibration massage modules. These treatment modules are adjusted in real time by the control module and work according to the personalized treatment parameters generated by the neural network model. The treatment system uses safe and non-invasive low-frequency modulated medium-frequency pulse technology, combined with controllable temperature hot compress, to provide a comfortable and effective treatment experience.

[0066] The rehabilitation therapy applet that comes with the wearable rehabilitation therapy device has a simple and easy-to-use interface. Users can obtain diagnostic parameters through diagnostic questionnaires, computer vision analysis, and body function data import. The wearable rehabilitation therapy device has built-in sensors to monitor the status of the corresponding parts (such as the waist) in real time. The data during the treatment process will be recorded, and part of it will be processed at any time at the wearable rehabilitation therapy device through edge computing and finally uploaded to the cloud for optimizing subsequent treatment plans or feedback on user treatment effects.

[0067] For the sake of convenience of explanation, the following examples are mostly used in this article for rehabilitation treatment of the waist. Such description should not be understood as any limitation to the present invention. Rehabilitation treatment of other body parts, such as shoulders, neck, knee joints, etc., can adopt the scheme disclosed in this invention, which is also included in the scope of patent protection of the present invention.

[0068] See also Figure 1, which shows a model training process according to an embodiment of the present disclosure, which is used to train a neural network model for generating treatment parameters for a user.

[0069] like Figure 1 As shown, in step S10, the neural network model is trained according to a sample data set, wherein each sample data in the sample data set is annotated with a first diagnostic parameter and a corresponding first treatment parameter; in step S20, the neural network model determines a corresponding second treatment parameter according to the user's second diagnostic parameter for use in the wearable rehabilitation therapy device; in step S30, the neural network model updates the second diagnostic parameter according to treatment feedback data of the user using the wearable rehabilitation therapy device for use in subsequent rehabilitation treatment of the user; in step S40, the neural network model is optimized according to the updated sample data set, and the sample data set is updated according to the treatment feedback data of one or more users using the wearable rehabilitation therapy device.

[0070] Here, the goal of neural network training is to learn the mapping relationship between input (diagnostic parameters) and output (treatment parameters). Specifically:

[0071] Diagnostic parameters are input features, such as pain type at the target site (such as low back pain type), pain intensity (VAS score), motor function (angle), muscle strength at the target site (such as waist muscle strength), etc.

[0072] The treatment parameters are the target outputs, such as electrical stimulation waveform, frequency, pulse width, heat compress temperature, etc.

[0073] Each record in the sample dataset represents a complete sample, containing a specific input-output pair, where input feature = diagnostic parameter; output target = corresponding treatment parameter. The sample dataset can be stored in the cloud for easy updating.

[0074] Each data sample may include the following input features:

[0075] Type of low back pain (categorical feature): such as "non-specific low back pain", "lumbar fasciitis", "chronic low back pain", "lumbar disc herniation". Use one-hot encoding or categorical labels (such as 0, 1, 2, 3) to represent.

[0076] Pain intensity (numerical feature): This is a user-subjective indicator using a visual analogue scale (VAS) ranging from 0 to 10 points.

[0077] Motor function (numerical features): objective indicators, including standing multi-joint flexion angle (unit: °), such as 90°-120°, standing extension angle (unit: °), such as 10°-30°. It can be obtained through motion test results captured by computer vision or clinical functional test reports.

[0078] Lumbar muscle strength (numerical feature): assessed by muscle strength score ranging from 0 to 5. This can be obtained through motion test results captured by computer vision or muscle strength sensors built into wearable rehabilitation therapy devices.

[0079] Core endurance time (numerical feature): (unit: seconds), for example, the core endurance test time is 20 seconds.

[0080] The output target (treatment parameter) is the parameter that the neural network needs to predict and is used to guide the treatment execution of the wearable device. For example, it may include:

[0081] Electrical stimulation waveform (categorical feature): such as "bi-phase symmetrical square wave" or "triangle wave". Use categorical labels or one-hot encoding.

[0082] Frequency range (numerical feature): including low frequency (1-200Hz), medium frequency modulation (6000-7000Hz). Record the frequency range or center frequency and store it as a continuous value.

[0083] Pulse width (numerical characteristics): such as 200 microseconds.

[0084] Hot compress temperature (numerical feature): such as 38-42°C, which can be adjusted dynamically according to the user's tolerance. Each session lasts 15 minutes.

[0085] Vibration massage intensity (numerical feature): For example, it is divided into 0-25 levels from no intensity to high intensity and is represented by discrete values.

[0086] Treatment mode (classification feature): such as "hot compress mode", "manipulation mode", "combined hot compress and electrical stimulation mode", etc. Use classification labels (such as 0, 1, 2) to represent.

[0087] Here, the key to establishing the mapping relationship is:

[0088] Acquisition of diagnostic parameters: Obtaining the patient's health status through questionnaires, measuring devices or video analysis.

[0089] Determine the treatment plan: First, a treatment plan is proposed through the neural network model based on the collected diagnostic parameters. Then, a doctor or sports rehabilitation expert will manually review and confirm the treatment parameters based on experience and guidelines.

[0090] Sample data can be annotated in the following ways:

[0091] 1) Expert annotation

[0092] The data are manually annotated by doctors or rehabilitation therapists based on patient records and treatment guidelines.

[0093] Annotated treatment parameters are based on known best practices, such as adjusting frequency and heat temperature based on pain intensity and motor function.

[0094] 2) Automatic labeling tool

[0095] Use rule engines such as those that generate treatment parameters based on thresholds and formulas.

[0096] Example: If the pain intensity is ≥8, the frequency is adjusted to 90 Hz and the hot compress temperature is increased to 40°C.

[0097] Return to Reference Figure 1 In step S10, the neural network model is trained according to the sample data set, and each sample data in the sample data set is marked with a first diagnostic parameter and a corresponding first treatment parameter.

[0098] Here, the neural network model can adopt a multi-layer perceptron (MLP) structure to process mixed types of input features and output multiple types of treatment parameters.

[0099] According to one example, the input feature, namely the first diagnostic parameter, includes:

[0100] Low back pain type → categorical feature (4 categories), dimension is 4 after one-hot encoding.

[0101] Pain intensity → Numerical feature, directly use the original value (range 0-10), which needs to be standardized.

[0102] Motor function → Numerical features, including flexion and extension angles, normalized to a range of 0 to 1.

[0103] Lumbar muscle strength → numerical characteristics, standardized.

[0104] Core endurance time → numerical feature, normalized.

[0105] The total input dimensions are:

[0106] Low back pain type (unique hot encoding, 4 dimensions) + pain intensity (1 dimension) + motor function (2 dimensions) + waist muscle strength (1 dimension) + core endurance time (1 dimension) = 9-dimensional input

[0107] Multiple hidden layers (fully connected layers) are selected to extract deep information of the input features. ReLU activation function can be used to introduce nonlinear relationships. The hidden layers are designed to gradually extract higher-level feature representations to ensure that the network can capture complex feature relationships.

[0108] Design the output layer according to different output goals:

[0109] Electrical stimulation waveform (classification feature): Use the Softmax activation function to output the number of categories (e.g., 2 waveform types).

[0110] Frequency range (numerical feature): Use a linear activation function to output a continuous value.

[0111] Pulse width (numerical feature): Use a linear activation function to output a continuous value (such as 200 microseconds).

[0112] Hot compress temperature (numerical feature): Use a linear activation function to output a continuous value (38-42°C).

[0113] Vibration massage intensity (numerical feature): Use a linear activation function to output a continuous value (such as 25 levels).

[0114] Treatment mode (classification feature): Use the Softmax activation function to output the number of categories (e.g. 2 modes).

[0115] During the training process, different loss functions are used for different tasks. For classification tasks, such as electrical stimulation waveform, vibration massage intensity, and treatment mode, the cross entropy function is used to measure the difference between the predicted probability output by the model and the true label. For regression tasks, such as frequency, pulse width, and hot compress temperature, the mean square error is used to measure the square difference between the model's predicted value and the true value, which is suitable for predicting continuous numerical data. When the loss function converges or reaches the preset number of training rounds (Epochs), the neural network model training is completed.

[0116] In some embodiments, the neural network model can also use a convolutional neural network (CNN), and if the first diagnostic parameter contains time series data or there is a dependency between the first diagnostic parameters (for example, the patient's medical history, different treatment stages), a recurrent neural network (RNN) or a long short-term memory network (LSTM) can capture these temporal dependencies. If the first diagnostic parameter comes from different data sources (such as text, images, sensor data, etc.), a multimodal network can be used. Each data source uses a dedicated network structure (such as CNN for images and MLP for numerical data), and the respective outputs are fused to finally predict the first treatment parameter. The specific model structure depends on the complexity of the input data.

[0117] In step S20, the trained neural network model determines corresponding second treatment parameters according to the user's second diagnostic parameters for use in the wearable rehabilitation treatment device.

[0118] According to some embodiments, the user can input the second diagnostic parameter through the rehabilitation therapy applet on his mobile phone. Among them, for the type of low back pain, the user can determine it through the rehabilitation therapy applet or through hospital diagnosis. For the intensity of pain, it is up to the user to make a subjective judgment. For motor function, waist muscle strength, core endurance time, etc., it can be evaluated through computer vision algorithms and built-in sensors of wearable rehabilitation therapy devices.

[0119] Motor function assessment is mainly achieved by detecting and analyzing the patient's joint range of motion and movement quality. According to an example, a camera is used to obtain a video of the user completing a specified action, such as standing flexion and extension, forward bending, backward bending, etc. Ensure good lighting, the patient is facing the camera directly or sideways, and the background is as simple as possible. Use a posture estimation algorithm to extract the coordinates of the user's skeletal key points, such as shoulders, hips, knees, and ankles, and form a skeleton model based on the extracted skeletal key point coordinates for motion analysis; calculate the flexion angle based on the skeletal key points, such as the knee and lumbar spine, and extract the maximum joint angle and minimum angle to calculate the range of motion, for example, the standing flexion angle range (90°-120°).

[0120] Lumbar muscle strength can be evaluated through movement performance and stability of posture control. According to one example, a camera is used to record specific movements of the user, such as the user's movements in a specific posture (such as bridge or plank support), or to record the user completing resistance movements (such as bending over to lift heavy objects). The posture estimation algorithm is used to extract bone key points to detect the relative angle and stability of the lumbar spine and pelvis. The waist muscle strength is scored based on the movement characteristics of the bone key points, such as 0-5 points; or the sensors on the wearable device are used to extract data such as the posture offset of the lumbar spine and pelvis, the amplitude of muscle jitter, and the accuracy of posture control jitter.

[0121] Core endurance time is usually measured by evaluating the ability of the core muscle groups to maintain a certain posture for a specific period of time. According to one example, core endurance tests usually include static support movements (such as plank support, side bridge, etc.), and the time the user maintains the posture can be monitored by visual algorithms. Use a camera to record the user's movements to complete the core endurance test, such as plank support, side bridge, etc. Identify whether the user has entered a standard core endurance posture (such as plank support). If the posture meets the preset standards (for example, the angle of the body straight line is close to horizontal), start timing. Continue to analyze video frames and monitor the user's posture. If the user's posture always remains within the standard range, continue timing. When it is detected that the user's posture deviates from the standard (such as the body bending or collapsing beyond the threshold), stop timing. This duration is the user's core endurance time. Alternatively, data can also be extracted through the sensors provided by the wearable device.

[0122] After the rehabilitation treatment applet obtains all the second diagnostic parameters, it passes them to the neural network model. Subsequently, the neural network model determines the corresponding second treatment parameters according to the second diagnostic parameters of the user, so as to be used for the rehabilitation treatment of the user by the wearable rehabilitation treatment device.

[0123] According to one example, the input second diagnostic parameters include:

[0124] Type of low back pain: chronic low back pain; pain intensity: 7; motor function: standing multi-joint flexion angle: 95°, standing extension angle: 15°; waist muscle strength: 3; core endurance time: 25 seconds.

[0125] The second treatment parameters output by the neural network model include:

[0126] Electrical stimulation waveform: two-phase symmetrical square wave; frequency range: low frequency 80Hz, medium frequency modulation 6500Hz; pulse width: 220 microseconds; hot compress temperature: 40℃; vibration massage intensity: medium; treatment mode: combined mode of hot compress and electrical stimulation.

[0127] The control module in the wearable rehabilitation therapy device activates the electrical stimulation module and the hot compress module at the same time according to the treatment mode parameters, such as the "hot compress and electrical stimulation combined mode", and sends each second treatment parameter to the corresponding treatment module. Each treatment module works according to the received parameters and runs for a specific time period according to the established plan to provide users with precise treatment. Specifically including:

[0128] Electrical stimulation module: Generates corresponding electrical pulse signals based on the received electrical stimulation waveform, frequency range and pulse width parameters, and accurately applies them to the user's target area through electrodes to achieve the effects of relieving pain, improving local circulation and enhancing muscle function.

[0129] Hot compress module: adjust the temperature of the graphene hot compress device to 40°C and maintain the set temperature for a set time (such as 15 minutes) to promote blood circulation, reduce inflammation and relax waist muscles.

[0130] Vibration massage module: According to the received vibration massage intensity parameters, the vibration amplitude and frequency are adjusted to achieve moderate relaxation and massage of the waist muscles, further relieve tension and enhance the therapeutic effect.

[0131] In step S30, the neural network model updates the second diagnostic parameter of the user according to the treatment feedback data of the user using the wearable rehabilitation therapy device for subsequent treatment of the user.

[0132] In some embodiments, the wearable rehabilitation therapy device also monitors the treatment process in real time, and obtains treatment feedback data of the user based on the second treatment parameter during the treatment process and after the treatment is completed.

[0133] According to one example, the electrodes of the electrical stimulation module may also be embedded with an electromyography sensor (sEMG) and a muscle force sensor, the electromyography sensor being used to monitor the muscle response intensity (the degree of contraction or relaxation), and the muscle force sensor being used to detect the user's muscle strength. The graphene layer of the hot compress module may also be embedded with a temperature sensor for monitoring the temperature distribution of the hot compress area, thereby dynamically adjusting the hot compress temperature to avoid overheating or uneven temperature. The vibration massage module may also be embedded with a skin pressure sensor for measuring the contact pressure between the vibration massage module and the skin to ensure that the vibration massage module acts on the target area and optimizes the massage effect. The control module of the wearable rehabilitation therapy device may dynamically adjust the relevant parameter settings in the current treatment process based on the real-time monitoring of these sensors.

[0134] In some examples, the wearable rehabilitation therapy device is also configured with an IMU (inertial measurement unit) to capture motion trajectory, in which the accelerometer (3-axis) and gyroscope (3-axis) capture the dynamic posture and motion trajectory of the placed area. For example, when used for lumbar rehabilitation therapy, the 6-axis IMU collects the dynamic posture and motion trajectory of the user's waist, and the wearable rehabilitation therapy device obtains the user's current motion function parameters, core endurance time, etc. through edge computing.

[0135] In some examples, treatment feedback data can be obtained through subjective evaluations and objective monitoring data of patients after treatment, including:

[0136] 1) Subjective evaluation: changes in pain intensity (VAS score before and after comparison); patient ratings of treatment comfort and effectiveness (e.g., 1-5 stars).

[0137] 2) Objective indicators: Improved motor function after treatment (changes in flexion angle and extension angle); improved core endurance (extended time); changes in muscle strength (through comparison of muscle strength scores); skin temperature or muscle response data after electrical stimulation.

[0138] These treatment feedback data can be used to update the corresponding second diagnostic parameters of the current user, and the updated second diagnostic parameters will be used for the user's next treatment. For example, when the user uses the wearable rehabilitation therapy device next time, the corresponding second treatment parameters will be generated based on the updated second diagnostic parameters, so as to always be synchronized with the user's current health status, thereby providing the user with a personalized treatment plan.

[0139] The user's treatment feedback data will be transmitted to the user's device through the communication module, such as the rehabilitation treatment applet, so that the user can view the current treatment effect. At the same time, the user's treatment feedback data will be uploaded to the cloud for storage, so that the user's neural network model can be used for the next treatment, and can also be used for the optimization of the neural network model in the future.

[0140] In step S40, the neural network model is optimized based on the updated sample data set, and the sample data set is updated based on the treatment feedback data of one or more users.

[0141] According to an example, treatment feedback data can be used as a supplementary feature of the diagnostic parameters to form new sample data. That is:

[0142] Input features = second diagnostic parameter + second treatment parameter + treatment feedback data

[0143] Output target = dynamically adjusted treatment parameters

[0144] Based on this, the neural network model is retrained or fine-tuned using new sample data to improve the model's ability to optimize specific diagnostic features. This step helps the model continuously improve the accuracy and effectiveness of personalized treatment in the long run.

[0145] The neural network not only learns the mapping between diagnostic parameters and initial treatment parameters, but also adjusts the optimized treatment parameters based on feedback.

[0146] By continuously monitoring the treatment effects (e.g., changes in pain intensity, improvements in motor function, etc.), these treatment feedback data can be used as new feedback inputs to regularly update the training data set. This process can be used to fine-tune the model, so that the neural network model can continuously optimize and improve the treatment plan during continuous use.

[0147] Here, a large number of user treatment feedback data can be used to regularly update the sample data set so that the neural network can be continuously optimized. This enables the model to more accurately provide different users with treatment plans that are suitable for them. In addition, during the treatment of a specific user, the model will also update the user's diagnostic parameters based on the current user's treatment feedback to provide each user with a personalized treatment plan in real time.

[0148] In the model re-update / optimization phase, the treatment effect indicator is introduced as the optimization target. For example, the loss function weight is designed so that samples with poor treatment effects have a greater impact on training.

[0149] The loss function can be adjusted to, for example: L = L1 + α * L2

[0150] Among them, L is the total loss function, and the goal of neural network training is to minimize this value;

[0151] L1 is the error loss of the second treatment parameter prediction, which is used to measure the difference between the second treatment parameter predicted by the model and the true value. It can be the cross entropy loss in the classification task or the mean square error (MSE) in the regression task.

[0152] L2 is the loss based on the treatment feedback data, which is used to measure the deviation between the actual treatment effect and the expected effect. It can be defined according to specific effect indicators, such as the improvement value of VAS score, the improvement degree of motor function test, etc.

[0153] α is an adjustment weight used to balance the importance of the two losses. It is usually based on the role of treatment effect feedback in the overall optimization.

[0154] Newly collected treatment feedback data is added in each round of model updating to gradually optimize the model, which is similar to online learning.

[0155] The wearable rehabilitation therapy device disclosed herein can have the following uses:

[0156] 1. Sports rehabilitation: A dedicated mode designed for athletes or fitness enthusiasts to relieve waist fatigue after exercise, repair minor injuries, and prevent potential waist strain problems.

[0157] 2. Office health management: Suitable for office workers who sit for long periods of time. The product can provide preventive intervention before low back pain occurs, help relieve muscle tension, improve posture, and improve work efficiency.

[0158] 3. Management of chronic low back pain in middle-aged and elderly people

[0159] In view of the characteristics of chronic low back pain in middle-aged and elderly people, a low-intensity, long-term treatment model is designed, combined with real-time monitoring to improve treatment efficacy and safety.

[0160] 4. Family health management: The product is portable and easy to use, suitable for sharing by multiple family members, and enables precise treatment through personalized diagnostic data.

[0161] The wearable rehabilitation treatment device disclosed herein may be used by:

[0162] 1. Common low back pain patients: including mild to moderate patients with low back discomfort caused by long-term sitting, poor posture, heavy physical labor, etc., are the most widely used group.

[0163] 2. Rehabilitation treatment institutions: Rehabilitation centers can use the product as auxiliary treatment equipment and combine it with professional physical therapy to provide patients with scientific, standardized and personalized treatment plans.

[0164] 3. Primary medical institutions: When primary medical resources are limited, this product can become a portable low back pain treatment device for primary medical institutions to meet the needs of a large number of patients.

[0165] 4. Insurance and health management companies: Products can be integrated into insurance or health management plans to provide users with innovative health services and improve user satisfaction with the added value of insurance.

[0166] 5. Smart health equipment manufacturers: can be linked with other health management devices (such as smart watches, posture correction devices, etc.) to form a complete health ecosystem and provide a full range of solutions for patients with low back pain.

[0167] The wearable rehabilitation therapy device disclosed herein is applicable to:

[0168] 1. Multi-body expansion: This technology can be expanded to other parts such as the neck, shoulders, hip and knee joints, limbs, etc., and corresponding treatment equipment can be developed.

[0169] 2. Telemedicine support: Combining IoT technology and remote monitoring functions, it can enable doctors to provide remote guidance and adjust treatment plans, which is particularly suitable for patients in remote areas or with limited mobility.

[0170] 3. Preventive health management: In addition to treatment, this product can be combined with the AI ​​health management platform to predict and intervene in lumbar diseases, and promote the transformation of the health management model from "treatment-oriented" to "prevention-oriented".

[0171] Through the expansion of these extended scenarios and user subjects, the wearable rehabilitation therapy device disclosed in the present invention is not only limited to a single low back pain treatment device, but can also build a multi-functional, multi-scenario health management ecosystem to provide intelligent rehabilitation and health services for a wider range of people.

[0172] It should be noted that the various embodiments of the present disclosure may be implemented in software and / or a combination of software and hardware, for example, may be implemented using an application specific integrated circuit (ASIC), a general purpose computer, or any other similar hardware device. In one embodiment, the software program of the various embodiments of the present disclosure may be executed by a processor to implement the steps or functions described above. Similarly, the software program of the various embodiments of the present disclosure (including related data structures) may be stored in a computer-readable recording medium, for example, a RAM memory, a magnetic or optical drive, or a floppy disk and the like. In addition, some steps or functions of the various embodiments of the present disclosure may be implemented using hardware, for example, as a circuit that cooperates with a processor to perform various steps or functions.

[0173] In addition, at least a portion of the embodiments of the present disclosure may be applied as a computer program product, such as a computer program instruction, which, when executed by a computing device, can call or provide the method and / or technical solution according to the embodiments of the present disclosure through the operation of the computing device. The program instruction for calling / providing the method of the embodiments of the present disclosure may be stored in a fixed or removable recording medium, and / or transmitted through a data stream in a broadcast or other signal-bearing medium, and / or stored in a working memory of a computing device that runs according to the program instruction.

[0174] For those skilled in the art, it is obvious that the embodiments of the present disclosure are not limited to the details of the above exemplary embodiments, and the embodiments of the present disclosure can be implemented in other specific forms without departing from the spirit or basic features of the embodiments of the present disclosure. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive, and the scope of the embodiments of the present disclosure is limited by the attached claims rather than the above description, so it is intended to include all changes within the meaning and scope of the equivalent elements of the claims in the embodiments of the present disclosure. Any figure mark in the claims should not be regarded as limiting the claims involved. In addition, it is obvious that the word "including" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any specific order.

Claims

1. A method for generating treatment parameters for a wearable rehabilitation therapy device through a neural network model, wherein: The method comprises the following steps: Training a neural network model according to a sample data set, wherein each sample data in the sample data set is labeled with a first diagnostic parameter and a corresponding first treatment parameter; Obtaining a second diagnostic parameter of the user; Inputting the second diagnostic parameter into the trained neural network model to obtain a corresponding second treatment parameter; transmitting the second treatment parameter to a corresponding treatment module of the wearable rehabilitation treatment device to perform rehabilitation treatment on the user; The method further comprises the following steps: According to the treatment feedback data of the user based on the second treatment parameter, the second treatment parameter is updated for use in the next rehabilitation treatment of the user.

2. The method according to claim 1, wherein: The method further comprises the following steps: The sample data set is updated according to the treatment feedback data of one or more users using the wearable rehabilitation therapy device, thereby optimizing the neural network model.

3. The method according to claim 2, wherein: The second diagnostic parameter, the second treatment parameter and the treatment feedback data constitute a set of new sample data to update the sample data set.

4. The method according to claim 1, wherein: The wearable rehabilitation therapy device is suitable for the rehabilitation treatment of musculoskeletal diseases, including the waist, shoulder and neck, hip and knee joints and limbs.

5. The method according to claim 1, wherein: The first diagnostic parameters include pain type, pain intensity, motor function, muscle strength of the target area, and core endurance time; the corresponding first treatment parameters include electrical stimulation waveform, frequency, pulse width, hot compress temperature, vibration massage intensity, and treatment mode.

6. The method according to claim 1, wherein: The pain type and pain intensity of the target site in the second diagnostic parameter are input by the user; The motor function, the muscle strength of the target part, and the core endurance time in the second diagnostic parameters are determined by collecting the motion trajectories of the user performing the corresponding designated actions.

7. A wearable rehabilitation therapy device, comprising a control module, a communication module, a neural network model and one or more treatment modules; wherein: The communication module is used to receive a second diagnostic parameter of the user and report treatment feedback data of the user based on the second treatment parameter; The neural network model is used to obtain a corresponding second treatment parameter according to the second diagnostic parameter, wherein the second diagnostic parameter is also updated according to treatment feedback data of the user using the wearable rehabilitation therapy device; wherein the neural network model is trained according to a sample data set, each sample data in the sample data set is annotated with a first diagnostic parameter and a corresponding first treatment parameter; and the neural network model is optimized according to the updated sample data set, wherein the sample data set is updated according to the treatment feedback data of one or more users using the wearable rehabilitation therapy device; The control module is used to transmit the second treatment parameter to the corresponding treatment module, so that each treatment module works according to its corresponding second treatment parameter.

8. The method according to claim 7, wherein: The second diagnostic parameter, the second treatment parameter and the treatment feedback data constitute a set of new sample data to update the sample data set.

9. The method according to claim 7, wherein: The wearable rehabilitation therapy device is also configured with a plurality of sensors, and at least part of the therapy feedback data is acquired through the sensors.

10. A method for generating treatment parameters for a wearable rehabilitation therapy device through a neural network model, wherein: The method comprises the following steps: Training the neural network model according to a sample data set, wherein each sample data in the sample data set is annotated with a first diagnostic parameter and a corresponding first treatment parameter; According to the second diagnostic parameter of the user, generating corresponding second treatment parameters through the neural network model for use in the wearable rehabilitation treatment device; updating the second diagnostic parameter according to the treatment feedback data of the user using the wearable rehabilitation therapy device, so as to be used by the neural network model to generate the corresponding second treatment parameter for the user next time; The method further comprises the following steps: The neural network model is optimized according to the updated sample data set, and the sample data set is updated according to treatment feedback data of one or more users using the wearable rehabilitation therapy device.

Images