DeepMVS: Learning Multi-view Stereopsis
Authors:
- Po-Han HuangPHPo-Han Huang
- Kevin MatzenKMKevin Matzen
- Johannes KopfJKJohannes Kopf
- Narendra AhujaNANarendra Ahuja
- Jia-Bin HuangJHJia-Bin Huang
Top Cited Papers
- 1 June 2018
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2821-2830
Abstract
We present DeepMVS, a deep convolutional neural network (ConvNet) for multi-view stereo reconstruction. Taking an arbitrary number of posed images as input, we first produce a set of plane-sweep volumes and use the proposed DeepMVS network to predict high-quality disparity maps. The key contributions that enable these results are (1) supervised pretraining on a photorealistic synthetic dataset, (2) an effective method for aggregating information across a set of unordered images, and (3) integrating multi-layer feature activations from the pre-trained VGG-19 network. We validate the efficacy of DeepMVS using the ETH3D Benchmark. Our results show that DeepMVS compares favorably against state-of-the-art conventional MVS algorithms and other ConvNet based methods, particularly for near-textureless regions and thin structures.Keywords
All Related Versions
This publication has 28 references indexed in Scilit:
- Learning-based view synthesis for light field camerasACM Transactions on Graphics, 2016
- FlowNet: Learning Optical Flow with Convolutional Networks2015
- Learning to compare image patches via convolutional neural networks2015
- Fully convolutional networks for semantic segmentation2015
- Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation2014
- SUN3D: A Database of Big Spaces Reconstructed Using SfM and Object Labels2013
- A benchmark for the evaluation of RGB-D SLAM systems2012
- High-quality shape from multi-view stereo and shading under general illumination2011
- Accurate, Dense, and Robust Multiview StereopsisIEEE Transactions on Pattern Analysis and Machine Intelligence, 2009
- ShadowCuts: Photometric Stereo with Shadows2007