Home Rehabilitation Movement Posture Recognition and Assessment System

Abstract

A home rehabilitation posture recognition and assessment system includes: a patient-end device for capturing rehabilitation images and displaying guidance and feedback; a cloud server platform comprising a posture recognition module, a motion analysis module, and an artificial intelligence assessment module. The posture recognition module can build a skeletal model, the motion analysis module can calculate limb angles and the number of rehabilitation movements, and the artificial intelligence assessment module can evaluate the accuracy of the movements in real time; and a medical personnel-end device for remotely receiving images and analysis results, tracking rehabilitation progress, and adjusting the plan. This system has remote operation, real-time feedback, and automated analysis capabilities, thereby reducing the burden on healthcare, improving rehabilitation efficiency, and providing a highly versatile and low-cost smart rehabilitation solution.

Claims

1. A posture recognition and assessment system for home rehabilitation movements, comprising: A patient-side device includes an image acquisition module and an interactive interface module. The image acquisition module is used to acquire continuous image data of the patient performing rehabilitation exercises at home. The interactive interface module is used to display rehabilitation instructions and feedback information. A cloud server platform is connected to the patient-side device via a network. The cloud server platform includes a posture recognition module, a motion analysis module, and an artificial intelligence assessment module. The posture recognition module is used to extract human joint points from continuous image data and build a skeletal model. The motion analysis module calculates based on the skeletal model. The AI ​​assessment module calculates joint angles and the number of movements. It is first trained through deep learning using standard rehabilitation movement data. When communicating with the patient's device, it can compare the deviation between the patient's rehabilitation movements and the standard posture in real time, and generate posture scores and posture suggestions for real-time feedback to the interactive interface module. A medical personnel device is connected to the cloud server platform through the network. The medical personnel device receives continuous image data, analysis data and posture scores provided by the cloud server platform, enabling medical personnel to remotely interpret, track and adjust rehabilitation plans.

2. The posture recognition and assessment system for home rehabilitation movements as described in claim 1, wherein the patient-end device and the medical personnel-end device are either computers, tablets, or mobile devices, and the image capturing module is a built-in or external network camera.

3. The posture recognition and assessment system for home rehabilitation movements as described in claim 2, wherein the image acquisition module acquires fixed-format images through an image input / output related function library, wherein the fixed-format images are a continuous RGB image stream with a resolution of 1280×720 and a refresh rate of 30fps.

4. The posture recognition and assessment system for home rehabilitation movements as described in claim 3, wherein each frame of image captured by the posture recognition module is fed into a pre-trained posture estimation model for human posture estimation, accurately identifying and extracting the patient's key joint points.

5. The posture recognition and assessment system for home rehabilitation movements as described in claim 1, wherein the movement analysis module calculates the angles of six major joints, including the shoulder, elbow, hip, and knee, through the vector inner product principle, and further derives the trunk tilt angle parameter to quantify the range of motion and compensation of rehabilitation movements.

6. The posture recognition and assessment system for home rehabilitation movements as described in claim 1, wherein the medical personnel terminal device includes a remote interaction module, which transmits control commands to the interactive interface module of the patient terminal device according to a preset schedule of the rehabilitation plan to remind the patient to perform designated rehabilitation movements. When the image capturing module is in the activated state, the remote interaction module can simultaneously send a communication request between the two parties, prompting the medical personnel terminal device and the patient terminal device to establish an instant audio-visual connection for the medical personnel to provide remote guidance and interactive monitoring.

7. The posture recognition and assessment system for home rehabilitation movements as described in claim 1, wherein the patient-end device is further coupled with at least one wearable device, which is placed at a specific location on the patient's body and provides a positioning reference when the posture recognition module performs joint point capture, thereby forming an anchoring effect in complex background environments.

8. The posture recognition and assessment system for home rehabilitation movements as described in claim 7, wherein the wearable device is an image-recognizable luminous accessory, a luminous patch, a headband, a wristband, a knee brace, a weightlifting object, or an assistive stick.

Images