SMURF: Self-Teaching Multi-Frame Unsupervised RAFT with Full-Image Warping
It addresses optical flow estimation for computer vision applications, offering a significant performance boost over existing unsupervised methods.
The paper tackles unsupervised learning of optical flow by introducing SMURF, which improves state-of-the-art performance by 36% to 40% over prior methods and outperforms some supervised approaches.
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.