Unsupervised Deep Learning for Optical Flow Estimation

doi:10.1609/aaai.v31i1.10723

Unsupervised Deep Learning for Optical Flow Estimation

2017-02-12

0

OA

AI

DOI：10.1609/aaai.v31i1.10723

原文链接

原文求助

摘要

En 中文

Recent work has shown that optical flow estimation can be formulated as a supervised learning problem. Moreover, convolutional networks have been successfully applied to this task. However, supervised flow learning is obfuscated by the shortage of labeled training data. As a consequence, existing methods have to turn to large synthetic datasets for easily computer generated ground truth. In this work, we explore if a deep network for flow estimation can be trained without supervision. Using image warping by the estimated flow, we devise a simple yet effective unsupervised method for learning optical flow, by directly minimizing photometric consistency. We demonstrate that a flow network can be trained from end-to-end using our unsupervised scheme. In some cases, our results come tantalizingly close to the performance of methods trained with full supervision.

Unsupervised Deep Learning for Optical Flow Estimation

摘要

期刊

机构

引用论文