126
click to vote
CVPR
2004
IEEE
16 years 5 months ago
2004
IEEE
In this paper we perform 3D face tracking on corrupted video sequences. We use a deformable model, combined with a predictive filter, to recover both the rigid transformations and...