Orthopedic patient activity posture image intelligent analysis method and system

By establishing a closed-loop processing chain for video stream acquisition and illumination normalization, key point detection, motion trajectory analysis, and fall risk prediction, the problem of real-time monitoring of patient activity status in orthopedic wards has been solved, achieving high-precision risk behavior identification and fall warning, and improving the safety management level of orthopedic wards.

CN122244957APending Publication Date: 2026-06-19FOURTH MILITARY MEDICAL UNIVERSITY
1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
FOURTH MILITARY MEDICAL UNIVERSITY
Filing Date
2026-05-19
Publication Date
2026-06-19

AI Technical Summary

⚠Technical Problem

Existing technologies lack the ability to estimate human posture and track movement trajectories in real time in orthopedic wards, cannot continuously perceive changes in patients' activity status, have not established an activity constraint model associated with the postoperative rehabilitation stage of orthopedics, lack a real-time fall risk prediction mechanism based on center of gravity displacement trajectory, and cannot form a closed-loop feedback mechanism to optimize the monitoring system.

⚗Method used

By acquiring video streams and performing illumination normalization preprocessing, key point detection and skeletal topology construction, motion trajectory analysis, risk action detection and fall risk prediction based on orthopedic constraint perception, an end-to-end closed-loop processing link is formed. Combined with adaptive image preprocessing, human posture estimation, temporal motion analysis, orthopedic constraint perception, and center of gravity trajectory-driven fall prediction, all-weather intelligent monitoring of orthopedic patients' activities is achieved.

🎯Benefits of technology

It enables 24/7 intelligent monitoring of orthopedic patients' activities, with a risk behavior identification accuracy rate of over 93% and an early warning response time of less than 3 seconds, significantly reducing the false negative rate and compensating for the time blind spots of manual rounds.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
Patent Text Reader

Abstract

This invention discloses an intelligent analysis method and system for images of orthopedic patients' activity postures, belonging to the fields of medical image intelligent analysis and computer vision technology. The method acquires video streams through multi-angle cameras in the ward and performs adaptive illumination normalization and foreground-background difference preprocessing. It uses a top-down architecture to detect key points of the human body and construct a skeletal topology model. It identifies activity patterns through a spatiotemporal graph convolutional network, establishes a post-orthopedic post-operative activity constraint model for risk action detection, predicts fall risk based on center of gravity displacement trajectory and posture stability index, and achieves graded early warning. The accuracy rate of risk behavior recognition is over 93%, and the early warning response time is controlled within 3 seconds.
Need to check novelty before this filing date? Find Prior Art