Next Article in Journal
Estimating Tree Position, Diameter at Breast Height, and Tree Height in Real-Time Using a Mobile Phone with RGB-D SLAM
Next Article in Special Issue
Automated Mapping of Woody Debris over Harvested Forest Plantations Using UAVs, High-Resolution Imagery, and Machine Learning
Previous Article in Journal
The Dependence of Sea SAR Image Distribution Parameters on Surface Wave Characteristics
Open AccessArticle

Sparse Cost Volume for Efficient Stereo Matching

1
Graduate School of Information Science and Electrical Engineering, Kyushu University, Fukuoka 819-0395, Japan
2
Library, Kyushu University, Fukuoka 819-0395, Japan
3
Faculty of Information Science and Electrical Engineering, Kyushu University, Fukuoka 819-0395, Japan
*
Author to whom correspondence should be addressed.
Remote Sens. 2018, 10(11), 1844; https://doi.org/10.3390/rs10111844
Received: 22 October 2018 / Revised: 14 November 2018 / Accepted: 15 November 2018 / Published: 20 November 2018
Stereo matching has been solved as a supervised learning task with convolutional neural network (CNN). However, CNN based approaches basically require huge memory use. In addition, it is still challenging to find correct correspondences between images at ill-posed dim and sensor noise regions. To solve these problems, we propose Sparse Cost Volume Net (SCV-Net) achieving high accuracy, low memory cost and fast computation. The idea of the cost volume for stereo matching was initially proposed in GC-Net. In our work, by making the cost volume compact and proposing an efficient similarity evaluation for the volume, we achieved faster stereo matching while improving the accuracy. Moreover, we propose to use weight normalization instead of commonly-used batch normalization for stereo matching tasks. This improves the robustness to not only sensor noises in images but also batch size in the training process. We evaluated our proposed network on the Scene Flow and KITTI 2015 datasets, its performance overall surpasses the GC-Net. Comparing with the GC-Net, our SCV-Net achieved to: (1) reduce 73.08 % GPU memory cost; (2) reduce 61.11 % processing time; (3) improve the 3PE from 2.87 % to 2.61 % on the KITTI 2015 dataset. View Full-Text
Keywords: stereo matching; deep learning; 3D vision stereo matching; deep learning; 3D vision
Show Figures

Graphical abstract

MDPI and ACS Style

Lu, C.; Uchiyama, H.; Thomas, D.; Shimada, A.; Taniguchi, R.-I. Sparse Cost Volume for Efficient Stereo Matching. Remote Sens. 2018, 10, 1844.

Show more citation formats Show less citations formats
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Back to TopTop