FlowStep3D: Model Unrolling for Self-Supervised Scene Flow Estimation

Yair Kittenplon, Yonina C. Eldar, Dan Raviv

CVPR 2021 pp. 4114-4123

doi:10.1109/CVPR46437.2021.00410 /cvpr/2021/kittenplon2021cvpr-flowstep3d/

Abstract

Estimating the 3D motion of points in a scene, known as scene flow, is a core problem in computer vision. Traditional learning-based methods designed to learn end-to-end 3D flow often suffer from poor generalization. Here we present a recurrent architecture that learns a single step of an unrolled iterative alignment procedure for refining scene flow predictions. Inspired by classical algorithms, we demonstrate iterative convergence toward the solution using strong regularization. The proposed method can handle sizeable temporal deformations and suggests a slimmer architecture than competitive all-to-all correlation approaches. Trained on FlyingThings3D synthetic data only, our network successfully generalizes to real scans, outperforming all existing methods by a large margin on the KITTI self-supervised benchmark.

PDF CVPR Semantic Scholar

Cite

Text

Kittenplon et al. "FlowStep3D: Model Unrolling for Self-Supervised Scene Flow Estimation." Conference on Computer Vision and Pattern Recognition, 2021. doi:10.1109/CVPR46437.2021.00410

Markdown

[Kittenplon et al. "FlowStep3D: Model Unrolling for Self-Supervised Scene Flow Estimation." Conference on Computer Vision and Pattern Recognition, 2021.](https://mlanthology.org/cvpr/2021/kittenplon2021cvpr-flowstep3d/) doi:10.1109/CVPR46437.2021.00410

BibTeX

@inproceedings{kittenplon2021cvpr-flowstep3d,
  title     = {{FlowStep3D: Model Unrolling for Self-Supervised Scene Flow Estimation}},
  author    = {Kittenplon, Yair and Eldar, Yonina C. and Raviv, Dan},
  booktitle = {Conference on Computer Vision and Pattern Recognition},
  year      = {2021},
  pages     = {4114-4123},
  doi       = {10.1109/CVPR46437.2021.00410},
  url       = {https://mlanthology.org/cvpr/2021/kittenplon2021cvpr-flowstep3d/}
}