Object Extraction by Spatio-Temporal Assembling

Xiaoke Qin,Liang Tang,Jie Zhou
DOI: https://doi.org/10.1109/ICIP.2007.4379788
2007-01-01
Abstract:Among various algorithms for vision-based traffic monitoring, spatio-temporal (ST) slice analysis is attractive by computing over a larger temporal scale. However, it is unsuitable for further pattern recognition, since the conventional ST slice cannot preserve the spatial relationship of the original object image. In this paper, we propose a novel algorithm for accurate traffic object extraction. Compared with previous ST algorithms depending on one line per frame, we assemble the object based on foreground strips obtained from each frame and carefully designed motion estimation. Thus, both the spatial and temporal information is used more effectively. Applications in real canal traffic scenes show the advantages of our algorithm.
What problem does this paper attempt to address?