SelFlow: Self-Supervised Learning of Optical Flow

CVPR 2019  ·  Pengpeng Liu, Michael Lyu, Irwin King, Jia Xu ·

We present a self-supervised learning approach for optical flow. Our method distills reliable flow estimations from non-occluded pixels, and uses these predictions as ground truth to learn optical flow for hallucinated occlusions. We further design a simple CNN to utilize temporal information from multiple frames for better flow estimation. These two principles lead to an approach that yields the best performance for unsupervised optical flow learning on the challenging benchmarks including MPI Sintel, KITTI 2012 and 2015. More notably, our self-supervised pre-trained model provides an excellent initialization for supervised fine-tuning. Our fine-tuned models achieve state-of-the-art results on all three datasets. At the time of writing, we achieve EPE=4.26 on the Sintel benchmark, outperforming all submitted methods.

PDF Abstract CVPR 2019 PDF CVPR 2019 Abstract

Datasets


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Optical Flow Estimation KITTI 2012 SelFlow Average End-Point Error 1.5 # 7
Optical Flow Estimation KITTI 2015 SelFlow Fl-all 8.42 # 12
Optical Flow Estimation Sintel-clean SelFlow Average End-Point Error 3.74 # 21
Optical Flow Estimation Sintel-final SelFlow Average End-Point Error 4.26 # 14

Methods


No methods listed for this paper. Add relevant methods here