STFlow: Self-Taught Optical Flow Estimation Using Pseudo Labels

doi:10.1109/TIP.2020.3024015

STFlow: Self-Taught Optical Flow Estimation Using Pseudo Labels

2020-01-01

14

OA

AI

DOI：10.1109/TIP.2020.3024015

Original

Save

View PDF

Abstract

En 中文

The Deep learning of optical flow has been an active area for its empirical success. For the difficulty of obtaining accurate dense correspondence labels, unsupervised learning of optical flow has drawn more and more attention, while the accuracy is still far from satisfaction. By holding the philosophy that better estimation models can be trained with better-approximated labels, which in turn can be obtained from better estimation models, we propose a self-taught learning framework to continually improve the accuracy using self-generated pseudo labels. The estimated optical flow is first filtered by bidirectional flow consistency validation and occlusion-aware dense labels are then generated by edge-aware interpolation from selected sparse matches. Moreover, by combining reconstruction loss with regression loss on the generated pseudo labels, the performance is further improved. The experimental results demonstrate that our models achieve state-of-the-art results among unsupervised methods on the public KITTI, MPI-Sintel and Flying Chairs datasets.

Keywords:

Optical imaging

Estimation

Optical variables control

Machine learning

Adaptive optics

Unsupervised learning

Optical fiber networks

Deep neural networks

optical flow

self-taught learning

unsupervised learning

STFlow: Self-Taught Optical Flow Estimation Using Pseudo Labels

Abstract

AI Summary

Journal

Organization