Abstract
A novel Temporal Motion Vector Filter (TF) is presented and evaluated for real-time object detection on compressed videos in MPEG-2, MPEG-4 or H.264/AVC formats. The filter significantly reduces the noisy motion vectors that do not represent a real object movement . The filter analyses the temporal coherence of block motion vectors to determine if they are likely to represent true motion in the recorded scene. Experiments are performed using the CLEAR metrics for object detection and public available video datasets from CAVIAR, PETS and CLEAR. These experiments demonstrate that the TF outperforms the Vector Median Filter, by providing better object detection accuracy with reduced computational complexity. The good results obtained by the TF make it suitable as a first step towards implementing systems that aim to detect and track objects from compressed video by using motion vectors. The TF could also be used to improve other techniques based on motion vectors such as Global Motion Estimation (GME) and Motion-Compensated Frame Interpolation (MCFI).
Highlights
O BJECT tracking techniques aims at tracking objects in consecutive video frames
This paper presents a novel Temporal Motion Vector Filter (TF) to remove noisy motion vectors for object tracking purpose with low computational effort
The CLEAR Multiple Object Detection metrics described in Kasturi et al work [16] were used to numerically compare the capability of the Vector Median Filter (VMF), Spatiotemporal Motion Vector Filter (STF), and the proposed Temporal Motion Vector Filter (TF) to correctly detect true objects motion
Summary
Despite the increasing microprocessors computational power in recent years, the processing required by object tracking techniques still consists in a bottleneck to their wider adoption, specially in low cost embedded equipment as surveillance cameras and mobile devices. To reduce this computational power demand, some techniques that extract object motion information from compressed video streams, instead of the raw video, have been developed
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call