A survey of advances in vision-based human motion capture and analysis
Computer Vision and Image Understanding - Special issue on modeling people: Vision-based understanding of a person's shape, appearance, movement, and behaviour
A Robust Video-Based Algorithm for Detecting Snow Movement in Traffic Scenes
Journal of Signal Processing Systems
Proceedings of the 6th international conference on Human-robot interaction
Hi-index | 0.00 |
Automatically understanding events happening at a site is theultimate goal of visual surveillance system. This paperinvestigates the challenges faced by automated surveillance systemsoperating in hostile conditions and demonstrates the developedalgorithms via a system that detects water crises within highlydynamic aquatic environments. An efficient segmentation algorithmbased on robust block-based background modeling andthresholding-with-hysteresis methodology enables swimmers to bereliably detected amid reflections, ripples, splashes and rapidlighting changes. Partial occlusions are resolved using a MarkovRandom Field framework that enhances the tracking capability of thesystem. Visual indicators of water crises are identified based onprofessional knowledge of water crises detection, based on which aset of swimmer descriptors has been defined. Through seamlesslyfusing the extracted swimmer descriptors based on a novelfunctional link network, the system achieves promising results forwater crises detection. The developed algorithms have beenincorporated into a live system with robust performance fordifferent hostile environments faced by an outdoor swimming pool.