SMURF: Self-Teaching Multi-Frame Unsupervised RAFT with Full-Image Warping

Austin Stone, Daniel Maurer, Alper Ayvaci, Anelia Angelova, Rico Jonschkowski
5/14/2021
cs.CV

Official code: google-research/google-research (39,000 stars, Apache-2.0, last commit 2026-04-30).

Abstract

We present SMURF, a method for unsupervised learning of optical flow that improves state of the art on all benchmarks by $36\%$ to $40\%$ (over the prior best method UFlow) and even outperforms several supervised approaches such as PWC-Net and FlowNet2. Our method integrates architecture improvements from supervised optical flow, i.e. the RAFT model, with new ideas for unsupervised learning that include a sequence-aware self-supervision loss, a technique for handling out-of-frame motion, and an approach for learning effectively from multi-frame video data while still only requiring two frames for inference.

View on arXivView PDF

Code Implementations(1)

google-research/google-researchOfficial100%
Apache-2.0

Google Research

39,0008,488C, ROct 4, 20185 months agoApache-2.0
aimachine-learningresearch

Similar Papers

Unsupervised Learning Optical Flow in Multi-frame Dynamic Environment Using Temporal Dynamic Modeling

89% match2023

Semi-Supervised Learning of Optical Flow by Flow Supervisor

86% match1 repo252022

Models Matter, So Does Training: An Empirical Study of CNNs for Optical Flow Estimation

86% match1 repo1.7k2018

SICAF: Time-Varying Focus Bottleneck for Self-Supervised Event-Based Optical Flow with Spiking Neural Network

86% match2026

EV-MGRFlowNet: Motion-Guided Recurrent Network for Unsupervised Event-based Optical Flow with Hybrid Motion-Compensation Loss

85% match2023

MEMFOF: High-Resolution Training for Memory-Efficient Multi-Frame Optical Flow Estimation

85% match1 repo1022025

Cite this paper

@article{stone2021smurf,
  title  = {SMURF: Self-Teaching Multi-Frame Unsupervised RAFT with Full-Image Warping},
  author = {Austin Stone and Daniel Maurer and Alper Ayvaci and Anelia Angelova and Rico Jonschkowski},
  year   = {2021},
  eprint = {2105.07014},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/2105.07014}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2105.07014.svg)](https://codeofpaper.com/paper/2105.07014)

Discussion