Patient restlessness escape behavior early warning method and system based on video image analysis
By constructing a multi-feature fusion analysis framework and a cross-camera tracking system in a closed psychiatric ward, the problem of the inability of existing technologies to comprehensively predict patients' agitated and escape behaviors was solved. This enabled multi-dimensional prediction and precise intervention of patient behavior, reducing the false alarm rate and false negative rate of the system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- DATONG SIXTH PEOPLES HOSPITAL (DATONG RONGFU MILITARY HOSPITAL DATONG MENTAL HEALTH CENTER DATONG SOCIAL WELFARE PSYCHIATRIC HOSPITAL)
- Filing Date
- 2026-03-18
- Publication Date
- 2026-06-02
AI Technical Summary
Existing technologies lack intelligent early warning methods for patient agitation and escape attempts that integrate multi-dimensional behavioral precursor features and have cross-camera full-domain tracking capabilities in closed psychiatric wards. As a result, they cannot effectively predict and intervene in patients' agitated and impulsive behaviors and escape attempts.
A multi-feature fusion analysis framework covering facial emotional state, skeletal posture and movement, and spatial movement trajectory is constructed. Cross-camera identity association and trajectory tracking are achieved through video image analysis. Combined with a time-series early warning mechanism, comprehensive risk assessment and hierarchical early warning are carried out, and a closed-loop feedback mechanism is established to optimize system parameters.
It significantly improves the perception dimension and warning comprehensiveness of behavioral precursors, reduces the false alarm rate and missed alarm rate, and enables early prediction and precise intervention of patients' agitated escape behavior.
Smart Images

Figure CN122135436A_ABST