Auto-Rectify Network for Unsupervised Indoor Depth Estimation
Date
2021
Authors
Bian, J.W.
Zhan, H.
Wang, N.
Chin, T.J.
Shen, C.
Reid, I.
Editors
Advisors
Journal Title
Journal ISSN
Volume Title
Type:
Journal article
Citation
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021; 44(12)
Statement of Responsibility
Jia-Wang Bian, Huangying Zhan, Naiyan Wang, Tat-Jun Chin, Chunhua Shen, and Ian Reid
Conference Name
Abstract
Single-View depth estimation using the CNNs trained from unlabelled videos has shown significant promise. However, excellent results have mostly been obtained in street-scene driving scenarios, and such methods often fail in other settings, particularly indoor videos taken by handheld devices. In this work,we establish that the complex ego-motions exhibited in handheld settings are a critical obstacle for learning depth. Our fundamental analysis suggests that the rotation behaves as noise during training, as opposed to the translation (baseline) which provides supervision signals. To address the challenge, we propose a data pre-processing method that rectifies training images by removing their relative rotations for effective learning. The significantly improved performance validates our motivation. Towards end-to-end learning without requiring pre-processing, we propose an Auto-Rectify Network with novel loss functions, which can automatically learn to rectify images during training. Consequently, our results out perform the previous unsupervised SOTA method by a large margin on the challenging NYUv2 dataset.We also demonstrate the generalization of our trained model in ScanNet and Make3D, and the universality of our proposed learning method on 7-Scenes and KITTI datasets.
School/Discipline
Dissertation Note
Provenance
Description
Access Status
Rights
© 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.