A two-stage scheme for multi-view human pose estimation

We present a two-stage scheme integrating voxel reconstruction and human motion tacking. By combining voxel reconstruction with human motion tracking interactively, our method can work in a cluttered background where perfect foreground silhouettes are hardly available. For each frame, a silhouette-based 3D volume reconstruction method and hierarchical tracking algorithm are applied in two stages. In the first stage, coarse reconstruction and tracking results are obtained, and then the refinement for reconstruction is applied in the second stage. The experimental results demonstrate our approach is promising. Although our method focuses on the problem of human body voxel reconstruction and motion tracking in this paper, our scheme can be used to reconstruct voxel data and infer the pose of many specified rigid and articulated objects.

[1]  Kiriakos N. Kutulakos,et al.  A Theory of Shape by Space Carving , 2000, International Journal of Computer Vision.

[2]  Qiang Chen,et al.  Robust 3D modeling from silhouette cues , 2009, 2009 IEEE International Conference on Acoustics, Speech and Signal Processing.

[3]  Yin Li,et al.  An Accelerated Human Motion Tracking System Based on Voxel Reconstruction under Complex Environments , 2009, ACCV.

[4]  Jian Liu,et al.  Markless Human Motion Capture by Voxel Labeling , 2009, 2009 Fifth International Conference on Image and Graphics.

[5]  Mohan M. Trivedi,et al.  Human Body Model Acquisition and Tracking Using Voxel Data , 2003, International Journal of Computer Vision.

[6]  W. Eric L. Grimson,et al.  Adaptive background mixture models for real-time tracking , 1999, Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149).

[7]  Pushmeet Kohli,et al.  PoseCut: Simultaneous Segmentation and 3D Pose Estimation of Humans Using Dynamic Graph-Cuts , 2006, ECCV.

[8]  Long Quan,et al.  A Surface Reconstruction Method Using Global Graph Cut Optimization , 2006, International Journal of Computer Vision.

[9]  Ankur Agarwal,et al.  3D human pose from silhouettes by relevance vector regression , 2004, Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004..

[10]  Luc Van Gool,et al.  Full body tracking from multiple views using stochastic sampling , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).