A Neural Network with Adversarial Loss for Light Field Synthesis from a Single Image

Simon Evain, Christine Guillemot

Abstract

This paper describes a lightweight neural network architecture with an adversarial loss for generating a full light field from one single image. The method is able to estimate disparity maps and automatically identify occluded regions from one single image thanks to a disparity confidence map based on forward-backward consistency checks. The disparity confidence map also controls the use of an adversarial loss for occlusion handling. The approach outperforms reference methods when trained and tested on light field data. Besides, we also designed the method so that it can efficiently generate a full light field from one single image, even when trained only on stereo data. This allows us to generalize our approach for view synthesis to more diverse data and semantics.

Download


Paper Citation


in Harvard Style

Evain S. and Guillemot C. (2021). A Neural Network with Adversarial Loss for Light Field Synthesis from a Single Image.In Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP, ISBN 978-989-758-488-6, pages 175-184. DOI: 10.5220/0010268701750184


in Bibtex Style

@conference{visapp21,
author={Simon Evain and Christine Guillemot},
title={A Neural Network with Adversarial Loss for Light Field Synthesis from a Single Image},
booktitle={Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP,},
year={2021},
pages={175-184},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010268701750184},
isbn={978-989-758-488-6},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP,
TI - A Neural Network with Adversarial Loss for Light Field Synthesis from a Single Image
SN - 978-989-758-488-6
AU - Evain S.
AU - Guillemot C.
PY - 2021
SP - 175
EP - 184
DO - 10.5220/0010268701750184