Non-Contact Patient Movement Monitoring via Video Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current patient movement monitoring methods, such as on-body wrist sensors and video analysis, are inadequate for detecting critical movements like delirium due to disturbances, incomplete body part capture, and environmental factors like lighting and coverings, leading to delayed detection of patient conditions.
Innovation Solution
A system utilizing video cameras, a motion unit, and a segmentation unit to continuously monitor patient movement by identifying clusters of motion and segmenting body parts in normal and darkened room conditions, without the need for physical sensors, and distinguishing patient movements from those of visitors or equipment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If on-body wrist sensors are used for patient movement monitoring, then movement detection capability is improved, but patient comfort and compliance deteriorate due to physical attachment requirements
Solution Approach 1:
The patent replaces mechanical body-worn sensors with an optical system using video cameras to capture patient movements. The camera-based system processes video frames to detect motion clusters and identify body part movements without requiring any physical contact with the patient, thereby eliminating the disturbance caused by wearable devices while maintaining movement detection capability
Solution Approach 2:
The system creates a visual copy of patient movements through video imaging. By capturing and analyzing video frames of the patient's body, the system reconstructs movement information optically rather than through mechanical sensors, allowing non-contact monitoring that preserves patient comfort and compliance
2Measurement precision
If video cameras are used for continuous patient monitoring, then comprehensive movement capture is improved, but system complexity increases due to environmental factor management
Solution Approach 1:
The patent extracts and isolates patient movement information from video frames by identifying motion clusters specific to the patient. The system separates patient-related motion from environmental changes by focusing on clustered motion patterns in temporal neighborhoods, effectively filtering out irrelevant environmental factors while capturing comprehensive patient movements
Solution Approach 2:
The system introduces motion clustering analysis as an intermediary processing step between raw video capture and movement interpretation. This intermediate layer groups pixels exhibiting similar motion patterns, serving as a mediator that distinguishes genuine patient movements from environmental artifacts without requiring complex environmental sensing
3Loss of energy
If sporadic observation by healthcare practitioners is used, then resource utilization is improved, but detection timeliness deteriorates due to observation gaps
Solution Approach 1:
The patent implements continuous automated video monitoring that operates without interruption, continuously capturing and analyzing patient movements. This replaces sporadic human observation with an uninterrupted automated system that maintains constant surveillance, eliminating detection gaps while requiring minimal human intervention for resource efficiency
4Measurement precision
If body part segmentation is performed to identify specific movements, then movement interpretation accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent segments the patient's body into distinct body parts by analyzing motion clusters in video frames. The system identifies and separates movements of different body regions (head, trunk, limbs) by clustering pixels with similar motion characteristics, enabling specific movement interpretation through spatial and temporal segmentation of motion data
Data Source
AI summary
A monitoring system (10) includes at least one video camera (14), a motion unit (40), and a segmentation unit (42). The at least one video camera (14) is configured to continuously receive video of a subject in normal and darkened room conditions. The motion unit (40) identifies clusters of motion of the subject based on respiratory and body part motion in the received video of the subject. The segmentation unit (42) segments body parts of the subject based on the identified clusters of subject motion.