Fast and accurate reconstruction of compressed color light field

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

20 Scopus citations

Abstract

Light field photography has been studied thoroughly in recent years. One of its drawbacks is the need for multi-lens in the imaging. To compensate that, compressed light field photography has been proposed to tackle the trade-offs between the spatial and angular resolutions. It obtains by only one lens, a compressed version of the regular multi-lens system. The acquisition system consists of a dedicated hardware followed by a decompression algorithm, which usually suffers from high computational time. In this work, we propose a computationally efficient neural network that recovers a high-quality color light field from a single coded image. Unlike previous works, we compress the color channels as well, removing the need for a CFA in the imaging system. Our approach outperforms existing solutions in terms of recovery quality and computational complexity. We propose also a neural network for depth map extraction based on the decompressed light field, which is trained in an unsupervised manner without the ground truth depth map.

Original languageEnglish
Title of host publicationIEEE International Conference on Computational Photography, ICCP 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1-11
Number of pages11
ISBN (Electronic)9781538625262
DOIs
StatePublished - 29 May 2018
Event2018 IEEE International Conference on Computational Photography, ICCP 2018 - Pittsburgh, United States
Duration: 4 May 20186 May 2018

Publication series

NameIEEE International Conference on Computational Photography, ICCP 2018

Conference

Conference2018 IEEE International Conference on Computational Photography, ICCP 2018
Country/TerritoryUnited States
CityPittsburgh
Period4/05/186/05/18

Fingerprint

Dive into the research topics of 'Fast and accurate reconstruction of compressed color light field'. Together they form a unique fingerprint.

Cite this