Dense Lagrangian motion estimation with occlusions

We couple occlusion modeling and multi-frame motion estimation to compute dense, temporally extended point trajectories in video with significant occlusions. Our approach combines robust spatial regularization with spatially and temporally global occlusion labeling in a variational, Lagrangian framework with subspace constraints. We track points even through ephemeral occlusions. Experiments demonstrate accuracy superior to the state of the art while tracking more points through more frames.

• Type: Journal
• Year: 2012
• Where: CVPR
• Authors: Susanna Ricco, Carlo Tomasi
• Keywords: Computer Vision, Cvpr 2012, Motion Estimation, Occlusion, Regularization