Monocular reconstruction of human translation in motion sequence by MTA
Li Zhang, Ling Li · 2007
A motion trend analysis method is proposed to reconstruct human motion from un-calibrated monocular video. In this paper, relocation of human subject is the main concern, which is one of the most difficult parts in motion reconstruction and critically affects the accuracy of human posture generation. Numerical studies are conducted on experiments from synthetic and real video data to demonstrate the correctness of the recovered human translation in the proposed system.