TRACKING OF EXTENDED SIZE TARGETS IN H.264 COMPRESSED VIDEO USING THE PROBABILISTIC DATA ASSOCIATION FILTER (TuePmOR3)

Author(s) : Vimal Thilak (New Mexico State University, USA) Charles Creusere (New Mexico State University, USA)

Abstract : Object detection and tracking play a significant role in critical applications such as video monitoring and remote surveillance. These systems employ compression to efficiently utilize the available bandwidth. An example of an efficient compression solution to low bit rate video applications is the recently proposed H.264/AVC video coding standard. In particular, H.264/AVC has been optimized for transmission over wireless channels making it an attractive candidate for use in remote surveillance systems. In this paper, we propose an algorithm that exploits motion vectors generated by the H.264 encoder for object detection and tracking. Experimental results demonstrate the effectiveness of the proposed method to detect and track objects in real video sequences.

Menu