Computing the Stereo Matching Cost With a Convolutional Neural Network

Jure Zbontar, Yann LeCun; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, pp. 1592-1599

Abstract


We present a method for extracting depth information from a rectified image pair. We train a convolutional neural network to predict how well two image patches match and use it to compute the stereo matching cost. The cost is refined by cross-based cost aggregation and semiglobal matching, followed by a left-right consistency check to eliminate errors in the occluded regions. Our stereo method achieves an error rate of 2.61 % on the KITTI stereo dataset and is currently (August 2014) the top performing method on this dataset.

Related Material


[pdf]
[bibtex]
@InProceedings{Zbontar_2015_CVPR,
author = {Zbontar, Jure and LeCun, Yann},
title = {Computing the Stereo Matching Cost With a Convolutional Neural Network},
booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2015}
}