CVPR 2007 – Bridging the Gap between Detection and Tracking for 3D Monocular Video-Based Motion Capture – A. Fossati, M. Dimitrijevic, V. Lepetit, P. Fua

Supplemental material

We present the results of the algorithm depicted in the paper, using as input video sequences of different subjects taken from very different viewpoints.

Sequence1

This sequence describes the whole algorithm and is structured as follows:

- Input video

- Detections Linked into the Complete Trajectory

- Initial 3D Poses Projected in the Images

- Synthesized Background Image

- Refined 3D poses

- Refined 3D poses seen from a different viewpoint

Sequence2 Sequence3

These sequences just show the input and output of our algorithm, and are structured as follows:

- Detections Linked into the Complete Trajectory

- Refined 3D poses

- Refined 3D poses seen from a different viewpoint