Camera-Based Subject Position Detection Without Wearable Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for monitoring the well-being of individuals in care facilities and private residences require wearable devices with accelerometers and buttons, necessitating interactivity that can be a barrier for those unable to engage with them.
Innovation Solution
A camera-based system using a trained artificial neural network (ANN) to automatically capture and analyze images of a room to detect the position and state of a subject, triggering alerts without the need for wearable devices, and continuously improving its accuracy through training with human data labelers and neural network back propagation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If wearable devices with accelerometers and buttons are used for monitoring, then emergency alerting can be triggered, but the system requires interactivity that creates barriers for occupants unable to engage with devices
Solution Approach 1:
The patent replaces the mechanical interaction system (accelerometers and buttons) with an optical detection system. A camera captures images of the occupant's environment, and image processing algorithms automatically detect falls and other emergency situations without requiring any physical interaction from the occupant. This substitution eliminates the interactivity barrier while maintaining reliable emergency alerting capability.
Solution Approach 2:
The monitoring system performs self-service by automatically detecting emergency situations through image analysis without requiring occupant participation. The system independently monitors the environment, identifies fall conditions, and triggers alerts without needing the occupant to wear or interact with any device, thereby resolving the contradiction between reliability and ease of operation.
2Measurement precision
If occupants must consistently wear wearable devices for monitoring, then position detection is possible, but this creates a barrier for those unable to engage with devices
Solution Approach 1:
The patent replaces the requirement for occupants to wear mechanical devices with an optical imaging system. The camera continuously captures images of the room environment, and image processing algorithms automatically determine occupant position and state without requiring the occupant to wear any device. This maintains measurement precision while eliminating the wearing requirement barrier.
Solution Approach 2:
The patent introduces an intermediary system (camera and image processing algorithms) that mediates between the monitoring objective and the occupant. Instead of directly requiring occupant engagement through wearable devices, the intermediary system captures visual information and processes it to infer position and state, thereby maintaining detection accuracy without the barrier of device wearing.
3Measurement precision
If image processing is performed continuously to detect subject position, then monitoring accuracy improves over time through training, but computational resources and processing time are consumed
Solution Approach 1:
The patent applies preliminary action through pre-training the machine learning model on extensive datasets before actual monitoring begins. The model is trained to recognize fall patterns and occupant positions in advance, so that during real-time monitoring, the system can quickly make accurate determinations without consuming excessive processing time. This preliminary training enables fast, accurate detection while managing computational resources.
Solution Approach 2:
The patent uses partial action by processing only the most critical information from each image frame. Instead of analyzing every pixel and detail, the system focuses image processing on identifying key features relevant to fall detection and position determination. This selective processing approach maintains high measurement precision while reducing the time and computational resources required for each analysis.
Data Source
AI summary
A system for detecting changes in the position of a subject is provided. An imaging device captures one or more images in a room where a subject may be located and provides the one or more images to an artificial neural network for processing. The ANN processes the images to determine a current state of a subject in the room. If the determined current state of the subject is associated with one or more alerts, the system delivers one or more alerts, which may be audible, visual, and/or haptic.