Self-supervised Depth Estimation based on Feature Sharing and Consistency Constraints

Julio Mendoza, Helio Pedrini

Abstract

In this work, we propose a self-supervised approach to depth estimation. Our method uses depth consistency to generate soft visibility mask that reduces the error contribution of inconsistent regions produced by occlusions. In addition, we allow the pose network to take advantage of the depth network representations to produce more accurate results. The experiments are conducted on the KITTI 2015 dataset. We analyze the effect of each component in the performance of the model and demonstrate that the consistency constraint and feature sharing can effectively improve our results. We show that our method is competitive when compared to the state of the art.

Download


Paper Citation


in Harvard Style

Mendoza J. and Pedrini H. (2020). Self-supervised Depth Estimation based on Feature Sharing and Consistency Constraints.In Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 5: VISAPP, ISBN 978-989-758-402-2, pages 134-141. DOI: 10.5220/0008975901340141


in Bibtex Style

@conference{visapp20,
author={Julio Mendoza and Helio Pedrini},
title={Self-supervised Depth Estimation based on Feature Sharing and Consistency Constraints},
booktitle={Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 5: VISAPP,},
year={2020},
pages={134-141},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0008975901340141},
isbn={978-989-758-402-2},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 5: VISAPP,
TI - Self-supervised Depth Estimation based on Feature Sharing and Consistency Constraints
SN - 978-989-758-402-2
AU - Mendoza J.
AU - Pedrini H.
PY - 2020
SP - 134
EP - 141
DO - 10.5220/0008975901340141