ML Anthology

Dense Lagrangian Motion Estimation with Occlusions

Susanna Ricco, Carlo Tomasi
CVPR 2012 pp. 1800-1807
doi:10.1109/CVPR.2012.6247877 /cvpr/2012/ricco2012cvpr-dense/

Abstract

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.

PDF CVPR Semantic Scholar

Cite

Text

Ricco and Tomasi. "Dense Lagrangian Motion Estimation with Occlusions." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012. doi:10.1109/CVPR.2012.6247877

Markdown

[Ricco and Tomasi. "Dense Lagrangian Motion Estimation with Occlusions." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012.](https://mlanthology.org/cvpr/2012/ricco2012cvpr-dense/) doi:10.1109/CVPR.2012.6247877

BibTeX

@inproceedings{ricco2012cvpr-dense,
  title     = {{Dense Lagrangian Motion Estimation with Occlusions}},
  author    = {Ricco, Susanna and Tomasi, Carlo},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2012},
  pages     = {1800-1807},
  doi       = {10.1109/CVPR.2012.6247877},
  url       = {https://mlanthology.org/cvpr/2012/ricco2012cvpr-dense/}
}