Lifelogging System Based on Averaged Hidden Markov Models: Dangerous Activities Recognition for Caregiver Support

In this paper a prototype lifelogging system for monitoring persons with cognitive disabilities and elderly people, as well as a method for automatic detection of dangerous activities are presented. The system allows remote monitoring of observed persons via Internet website and respects the privacy of the persons by displaying their silhouettes instead of actual images. Application allows viewing of both real-time and historic data. Lifelogging data (skeleton coordinates) needed for posture and activity recognition are acquired using Microsoft Kinect 2.0. Several activities are marked as potentially dangerous and generate alarms sent to the caregivers upon detection. Recognition models are developed using Averaged Hidden Markov Models with multiple learning sequences. Action recognition includes methods for differentiation between normal and potentially dangerous activities e.g. self-aggressive autistic behavior) using the same motion trajectory. Some activity recognition examples and results are presented.