Skip to main content

Learning Single-Image Depth From Videos Using Quality Assessment Networks

Author(s): Chen, Weifeng; Qian, Shengyi; Deng, Jia

Download
To refer to this page use: http://arks.princeton.edu/ark:/88435/pr19r72
Abstract: Depth estimation from a single image in the wild remains a challenging problem. One main obstacle is the lack of high-quality training data for images in the wild. In this paper we propose a method to automatically generate such data through Structure-from-Motion (SfM) on Internet videos. The core of this method is a Quality Assessment Network that identifies high-quality reconstructions obtained from SfM. Using this method, we collect single-view depth training data from a large number of YouTube videos and construct a new dataset called YouTube3D. Experiments show that YouTube3D is useful in training depth estimation networks and advances the state of the art of single-view depth estimation in the wild.
Publication Date: 2019
Citation: Chen, Weifeng, Shengyi Qian, and Jia Deng. "Learning Single-Image Depth From Videos Using Quality Assessment Networks." IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2019): pp. 5597-5606. doi:10.1109/CVPR.2019.00575
DOI: 10.1109/CVPR.2019.00575
ISSN: 1063-6919
EISSN: 2575-7075
Pages: 5597 - 5606
Type of Material: Conference Article
Journal/Proceeding Title: IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Version: Author's manuscript



Items in OAR@Princeton are protected by copyright, with all rights reserved, unless otherwise indicated.