Skip to main content

Deep Depth Completion of a Single RGB-D Image

Author(s): Zhang, Yinda; Funkhouser, Thomas

Download
To refer to this page use: http://arks.princeton.edu/ark:/88435/pr1rg11
Abstract: The goal of our work is to complete the depth channel of an RGB-D image. Commodity-grade depth cameras often fail to sense depth for shiny, bright, transparent, and distant surfaces. To address this problem, we train a deep network that takes an RGB image as input and predicts dense surface normals and occlusion boundaries. Those predictions are then combined with raw depth observations provided by the RGB-D camera to solve for depths for all pixels, including those missing in the original observation. This method was chosen over others (e.g., inpainting depths directly) as the result of extensive experiments with a new depth completion benchmark dataset, where holes are filled in training data through the rendering of surface reconstructions created from multiview RGB-D scans. Experiments with different network inputs, depth representations, loss functions, optimization methods, inpainting methods, and deep depth estimation networks show that our proposed approach provides better depth completions than these alternatives.
Publication Date: 2018
Citation: Zhang, Yinda, and Thomas Funkhouser. "Deep Depth Completion of a Single RGB-D Image." In IEEE/CVF Conference on Computer Vision and Pattern Recognition (2018): pp. 175-185. doi:10.1109/CVPR.2018.00026
DOI: 10.1109/CVPR.2018.00026
ISSN: 1063-6919
EISSN: 2575-7075
Pages: 175 - 185
Type of Material: Conference Article
Journal/Proceeding Title: IEEE/CVF Conference on Computer Vision and Pattern Recognition
Version: Author's manuscript



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