We present a new method to robustly and efficiently analyze foreground when we detect background for a fixed camera view by using mixture of Gaussians models and multiple cues. Th...
In this paper, an object-based video retrieval methodology for search in large, heterogeneous video collections is presented. The proposed approach employs a real-time, compressed-...
Vasileios Mezaris, Ioannis Kompatsiaris, Michael G...
This paper describes a vision based pedestrian detection and tracking system which is able to count people in very crowded situations like escalator entrances in underground stati...
This paper presents a new technique to reduce the storage cost of high quality 3D video. In 3D video [12], a sequence of 3D objects represents scenes in motion. Every frame is com...
— Intestinal motility assessment with video capsule endoscopy arises as a novel and challenging clinical fieldwork. This technique is based on the analysis of the patterns of in...