Patient Movement Detection Using Environmental Models and AI Classifiers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Healthcare providers face challenges in monitoring patients' movements and detecting abnormal activities in real-time, particularly in reducing costs and ensuring quality care, due to limited staff and increased regulatory demands, while also needing to provide accountability and access to patient information.
Innovation Solution
A system and method for patient movement detection and fall monitoring using an environmental model and a classifier network to track patient movements, classify normal or abnormal activities, and generate alerts, which includes a computing device with a processor and non-volatile storage, and a monitoring server that processes video and spectrogram data to detect falls and other events, sending notifications to designated staff.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual patient monitoring is used, then staff can provide personalized care, but labor costs increase and coverage is limited
Solution Approach 1:
The patent replaces manual mechanical monitoring with an automated electronic monitoring system using cameras, sensors, and AI algorithms. The system captures video feeds, processes them through machine learning models to detect falls and abnormal behaviors, and generates alerts automatically, eliminating the need for continuous manual observation while expanding monitoring capacity.
Solution Approach 2:
The system creates digital copies of patient environments through camera feeds and sensor data, allowing virtual monitoring of multiple patients simultaneously. The AI system analyzes these digital representations to detect events, enabling one staff member to effectively monitor many patients through the automated system.
2Reliability
If continuous video monitoring is implemented, then real-time fall detection is improved, but data storage requirements and processing complexity increase
Solution Approach 1:
The system extracts only the essential visual features needed for fall detection from complete video feeds. Instead of storing and analyzing entire video streams, the AI algorithm processes video frames to extract motion patterns, posture changes, and spatial relationships, retaining only the critical information needed for accurate fall detection while minimizing data storage requirements.
Solution Approach 2:
The system performs preliminary processing of video data by pre-training AI models on large datasets of normal and abnormal behaviors. This preliminary action enables the system to quickly classify new video inputs without requiring extensive real-time computation or storage of raw video data, as the decision-making framework is already established through prior training.
3Measurement precision
If AI classification algorithms are used, then movement classification accuracy is improved, but computational processing time increases
Solution Approach 1:
The system applies partial processing by focusing AI computation only on frames or regions containing potential events. Instead of analyzing every frame in detail, the system uses preliminary motion detection to identify areas of interest, then applies full AI classification only to those specific regions, reducing overall processing time while maintaining accuracy for actual events.
Data Source
AI summary
A system and method for patient movement detection and fall monitoring to address the need to proactively monitor patients to detect abnormal movements, in-room activity, and other movements associated with providing in-room care. The system comprises an environmental model which can be used to track the position and movement of a patient and a classifier network configured to receive movement data and classify a patient's movement as normal or abnormal movement. In addition to monitoring in-room activity, the system and method create safe zones within the room to ensure patients are proactively monitor in the event of a seizure, fall, or other unintended activity. The system will record and store in-room video in a secure environment. Videos and notifications are automatically sent to designated staff as events occur.