CN-121506508-B - Tracking method and system applied to rehabilitation process of orthopedic patient

CN121506508BCN 121506508 BCN121506508 BCN 121506508BCN-121506508-B

Abstract

The invention relates to the technical field of computers and discloses a tracking method and a tracking system applied to the rehabilitation process of an orthopedic patient, wherein the method comprises the steps of synchronously acquiring an image sequence and echo signals of the patient under free movement through high-frame-rate optical imaging and millimeter wave radar; the method comprises the steps of carrying out three-dimensional reconstruction of key points of a human body and extraction of micro-vibration and micro-Doppler characteristics of muscles, inputting joint tracks and muscle activation signals into a biomechanical constraint time sequence analysis model, outputting standardized joint angle, muscle strength and movement smoothness sequences, constructing a multi-dimensional rehabilitation process track diagram and generating a structured evaluation report. The system comprises a multi-mode sensing layer, a data preprocessing layer, a biomechanical analysis layer and a rehabilitation evaluation layer. The invention realizes non-sensitive, continuous and fine-granularity rehabilitation tracking, and remarkably improves the comprehensiveness, timeliness and individuation level of evaluation.

Inventors

Song Benjing
SHI XIAOTAO
HAN YAN
YE LIJUAN
LI SHIHONG
CHEN SONG
WANG WEI

Assignees

中国人民解放军西部战区总医院

Dates

Publication Date: 20260512
Application Date: 20260113

Claims (7)

1. A tracking method applied to rehabilitation of an orthopedic patient, comprising: The system comprises a high-frame-rate optical imaging unit and a millimeter wave radar sensing unit which are deployed in a home environment of a patient, and an optical image sequence and millimeter wave echo signals of the patient in a free activity state are synchronously acquired; Performing three-dimensional reconstruction on the key points of the human body on the optical image sequence to obtain a space coordinate time sequence track comprising six joints of shoulder, elbow, wrist, hip, knee and ankle; Extracting micro Doppler characteristics of the millimeter wave echo signals, analyzing micro vibration frequency spectrums of muscle tissues in the movement process, and further deducing a time function of a muscle activation state, wherein the method comprises the following steps: performing distance-Doppler conversion on the original millimeter wave echo signals to generate a distance-speed matrix; Extracting a speed spectrum of a corresponding region within a preset distance threshold between the trunk and the limbs of the human body; performing short-time Fourier transform on the velocity spectrum to obtain time-frequency distribution; identifying a low frequency component in the distribution having a frequency shift of less than 5 hz, the component corresponding to micro-vibrations produced by contraction of muscle fibers; calculating an energy integrated value of the low frequency component as a quantization index of the muscle activation intensity; inputting the time function of the joint space coordinate time sequence track and the muscle activation state to a preset biomechanical constraint time sequence analysis model, wherein the model performs physical consistency check and noise suppression on the original perception data based on a human body kinematics chain structure and a muscle-bone coupling dynamics equation, and outputs a standardized joint Qu Shenjiao degree sequence, a muscle activation strength sequence and a movement smoothness index, and the method comprises the following steps: Inputting the normalized joint three-dimensional coordinate sequence and the muscle activation intensity sequence into a cyclic neural network embedded in an anthropomorphic chain topological structure; explicitly introducing a coupling constraint term of the joint angle change rate and the adjacent joint motion and a linear relation constraint term between the muscle activation strength and the joint moment in the hidden state updating process; performing end-to-end optimization through a composite loss function comprising joint track prediction errors, muscle-joint dynamics consistency errors and motion smoothness regularization items, and outputting a standardized function recovery sequence; The joint track prediction error in the composite loss function adopts a mean square error measurement to predict the difference between joint coordinates and input coordinates, and the muscle-joint dynamics consistency error is defined as the Euclidean distance between the predicted joint moment and the moment obtained by the reverse thrust of the muscle activation intensity; Based on the standardized sequence and the index, a multi-dimensional rehabilitation process track diagram taking time as a horizontal axis and taking a function recovery dimension as a vertical axis is constructed, and the current rehabilitation stage is automatically divided according to clinical rehabilitation stage standards, so that a structured evaluation report comprising function improvement rate, symmetry deviation and compensation behavior identification is generated.
2. The tracking method applied to the rehabilitation process of the orthopedic patient according to claim 1, wherein the three-dimensional reconstruction of the key points of the human body is performed on the optical image sequence to obtain a space coordinate time sequence track comprising six joints of shoulder, elbow, wrist, hip, knee and ankle, and the method comprises the following steps: human body segmentation is carried out on each frame of optical image, and a front Jing Renti mask is extracted; performing thermodynamic diagram regression on pixels in the mask by using a convolutional neural network, and positioning two-dimensional key points; mapping the two-dimensional key points into three-dimensional coordinates under a world coordinate system by a triangulation method in combination with binocular or multi-visual difference information; And (3) applying Kalman filtering to the three-dimensional coordinate sequence to eliminate jump noise caused by shielding or illumination mutation, so as to form a smooth joint track.
3. The method for tracking the rehabilitation progress of the orthopedic patient according to claim 2, wherein constructing the multi-dimensional rehabilitation progress trace graph with time as a horizontal axis and functional recovery dimension as a vertical axis comprises: The moving average value of the maximum flexion and extension angles of each day is used for representing the joint movement range; The dynamic time regular distance of the angle track of the joint with the same name on the affected side and the healthy side under the same task is used for representing the motion symmetry; by detecting abnormal activation patterns of non-target joints when performing a specified action, the duration duty cycle thereof exceeding the baseline threshold is calculated as a compensation index.
4. The method of claim 3, wherein the structured assessment report is automatically generated at a time granularity of three days, weeks, and months, and supports a longitudinal comparison with the historical assessment results.
5. The tracking method applied to the rehabilitation process of the orthopedic patient according to claim 4 is characterized in that the high-frame-rate optical imaging unit consists of at least two infrared light supplementing cameras, and the millimeter wave radar sensing unit adopts a frequency modulation continuous wave system.
6. The tracking method applied to the rehabilitation process of the orthopedic patient according to claim 5, wherein the high-frame-rate optical imaging unit and the millimeter wave radar sensing unit are connected to a local edge computing node through a gigabit Ethernet, a time synchronization module is built in the node, and microsecond-level hardware-level time stamp alignment is achieved through a precise time protocol.
7. A tracking system for use in an orthopedic patient rehabilitation session, wherein the tracking of the orthopedic patient rehabilitation session is achieved using the method of any of claims 1 to 6, the system comprising: the multimode sensing layer is used for deploying the high-frame-rate optical imaging unit and the millimeter wave radar sensing unit and synchronously collecting the optical image sequence and the millimeter wave echo signals; the data preprocessing layer is used for carrying out three-dimensional reconstruction of key points of the human body on the optical image sequence and carrying out micro Doppler feature extraction on millimeter wave echo signals; The biomechanical analysis layer is used for inputting the reconstructed key point track and the extracted muscle activation signals into the biomechanical constraint time sequence analysis model and outputting a standardized function recovery sequence; And the rehabilitation evaluation layer is used for constructing a multidimensional rehabilitation process track diagram based on the standardized sequence, dividing rehabilitation stages and generating a structured evaluation report.

Description

Tracking method and system applied to rehabilitation process of orthopedic patient Technical Field The invention belongs to the technical field of computers, and particularly relates to a tracking method and a tracking system applied to rehabilitation processes of orthopedic patients. Background With the aging of population and frequent occurrence of sports injury, orthopedics rehabilitation has become an important link in clinical medical systems. Modern rehabilitation medicine emphasizes individualized, dynamic and data-driven intervention strategies, the core of which is the continuous quantitative assessment of patient joint activity, muscle stress patterns and daily functional actions. However, the current mainstream rehabilitation monitoring means still highly depend on periodic clinic follow-up visit or video-based action analysis systems, the former can only acquire static indexes at discrete time points and cannot reflect the real function recovery state in the home environment, and the latter is limited by fixed scene and equipment deployment, so that all-weather and non-sensible natural behavior capture is difficult to realize. In addition, a part of wearable schemes adopt a rigid Inertial Measurement Unit (IMU), and although the IMU has certain continuous monitoring capability, key biomechanical signals such as micro-joint flexion and extension, muscle group cooperative activation and the like are distorted or lost due to heavy volume, uncomfortable wearing and easy skin slippage interference, so that the accuracy and timeliness of rehabilitation process judgment are seriously affected. The accurate rehabilitation tracking for orthopaedics patients is in need of a sensing mechanism capable of seamlessly fitting human bodies and sensing fine physiological actions with high sensitivity. The sensing mechanism can stably capture low-amplitude high-frequency signals such as 0.5 degree micro-flexion of knee joints, ankle joint dorsiflexion moment change or local myoelectric activation time sequence and the like, and keeps long-term wearing comfort and signal reliability in daily activities. The flexible electronic technology provides brand new possibility for constructing the biomechanical sensing interface by virtue of the skin-like mechanical property and high ductility, but how to effectively couple the biomechanical sensing interface with clinical evaluation standards of rehabilitation medicine is still lack of systematic solutions. Whether based on optical dynamic capturing, pressure insoles or traditional IMU rehabilitation monitoring systems, the problems of single sensing dimension, insufficient spatial resolution, poor user compliance and the like generally exist. Especially in the home or community rehabilitation scene, the patient has small action amplitude, low frequency and unstructured, the existing equipment is difficult to distinguish therapeutic inching and random limb shaking, and the multi-joint coupling movement and muscle response state cannot be analyzed synchronously. This results in a failure of the rehabilitation practitioner to recognize timely compensatory actions, motor inhibition or early signs of functional deterioration, missing the intervention window. Disclosure of Invention The invention provides a tracking method and a tracking system applied to the rehabilitation progress of an orthopedic patient, and aims to solve the technical problems that the traditional rehabilitation monitoring relies on an inertial sensor which is inconvenient to evaluate or wear by a doctor at a discrete time point, and daily fine actions (such as joint micro-flexion and muscle activation states) are difficult to continuously and accurately capture, so that the rehabilitation progress evaluation is incomplete. According to the invention, a non-contact multi-mode perception fusion architecture is constructed, high-frame-rate optical imaging and millimeter wave radar signals are combined, all-weather and non-inductive acquisition is carried out on limb movement of a patient in a natural living scene, a time sequence action analysis model based on biomechanical constraint is introduced, and fine granularity quantification on joint angle change, muscle activation intensity and movement coordination is realized, so that dynamic, continuous and traceable rehabilitation progress digital figures are generated. The invention provides a tracking method applied to rehabilitation process of an orthopedic patient, which comprises the following steps: The system comprises a high-frame-rate optical imaging unit and a millimeter wave radar sensing unit which are deployed in a home environment of a patient, and an optical image sequence and millimeter wave echo signals of the patient in a free activity state are synchronously acquired; Performing three-dimensional reconstruction on the key points of the human body on the optical image sequence to obtain a space coordinate time sequence track comprising six j