New Paper on Real-time Object Detection and Tracking
My paper, titled as "Real-time Detection and Tracking of Multiple Objects with Partial Decoding in H.264|AVC Bitstream Domain", was submitted to SPIE Electronic Imaging at June 2008 and was finally accepted at September 2008. It is an outcome of the research project which was performed at MCCB Lab in KAIST-ICC (ICU) during 2007. This paper will be orally presented by one of authors at the 21st Annual IS&T/SPIE Symposium on Electronic Imaging at January 19, 2009.
The novelty of the proposed algorithm is not only that it can detect and track all moving objects fast by using encoded information of H.264|AVC compressed videos, but also that it shows reliable performance even in complex scenes such as occlusion which previous algorithms have not dealt with. The following videos are the results of detecting and tracking pedestrians indoors and outdoors on the basis of the proposed algorithm.