Lightless Fields: Enhancement and Denoising of Light-Deficient Light Fields

Carson Vogt*, Geng Lyu, Kartic Subr

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Modern focused light field cameras are capable of capturing video at over 160 frames per second, but in so doing sacrifice shutter speed. Outside of laboratory environments, lighting can be problematic resulting in noisy light fields and poor depth reconstruction. To enhance and denoise modern focused light field cameras, we create a unique deep neural network that allows for the full light field to be processed at once, eliminates stitching artifacts, and takes advantage of feature redundancy between neighboring microlenses. We show that our double U-Net network, ENH-W, significantly outperforms several popular architectures and light field denoising methods in both visual and depth metrics.

Original languageEnglish
Title of host publicationAdvances in Visual Computing. ISVC 2020
PublisherSpringer
Pages383-396
Number of pages14
ISBN (Electronic)9783030645564
ISBN (Print)9783030645557
DOIs
Publication statusPublished - 7 Dec 2020
Event15th International Symposium on Visual Computing 2020 - San Diego, United States
Duration: 5 Oct 20207 Oct 2020

Publication series

NameLecture Notes in Computer Science
Volume12509
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference15th International Symposium on Visual Computing 2020
Abbreviated titleISVC 2020
Country/TerritoryUnited States
CitySan Diego
Period5/10/207/10/20

Keywords

  • Deep learning
  • Focused light field camera
  • Low light

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Lightless Fields: Enhancement and Denoising of Light-Deficient Light Fields'. Together they form a unique fingerprint.

Cite this