Separable Flow: Learning Motion Cost Volumes for Optical Flow Estimation

Feihu Zhang, Oliver J. Woodford, Victor Adrian Prisacariu, Philip H.S. Torr

ICCV 2021 pp. 10807-10817

doi:10.1109/ICCV48922.2021.01063 /iccv/2021/zhang2021iccv-separable/

Abstract

Full-motion cost volumes play a central role in current state-of-the-art optical flow methods. However, constructed using simple feature correlations, they lack the ability to encapsulate prior, or even non-local, knowledge. This creates artifacts in poorly constrained, ambiguous regions, such as occluded and textureless areas. We propose a separable cost volume module, a drop-in replacement to correlation cost volumes, that uses non-local aggregation layers to exploit global context cues and prior knowledge, in order to disambiguate motions in these regions. Our method leads both the now standard Sintel and KITTI optical flow benchmarks in terms of accuracy, and is also shown to generalize better from synthetic to real data.

PDF ICCV Semantic Scholar

Cite

Text

Zhang et al. "Separable Flow: Learning Motion Cost Volumes for Optical Flow Estimation." International Conference on Computer Vision, 2021. doi:10.1109/ICCV48922.2021.01063

Markdown

[Zhang et al. "Separable Flow: Learning Motion Cost Volumes for Optical Flow Estimation." International Conference on Computer Vision, 2021.](https://mlanthology.org/iccv/2021/zhang2021iccv-separable/) doi:10.1109/ICCV48922.2021.01063

BibTeX

@inproceedings{zhang2021iccv-separable,
  title     = {{Separable Flow: Learning Motion Cost Volumes for Optical Flow Estimation}},
  author    = {Zhang, Feihu and Woodford, Oliver J. and Prisacariu, Victor Adrian and Torr, Philip H.S.},
  booktitle = {International Conference on Computer Vision},
  year      = {2021},
  pages     = {10807-10817},
  doi       = {10.1109/ICCV48922.2021.01063},
  url       = {https://mlanthology.org/iccv/2021/zhang2021iccv-separable/}
}