Efficient high dynamic range video using multi-exposure CNN flow

  • Yuchen Guo
  • , Zhifeng Xie*
  • , Wenjun Zhang
  • , Lizhuang Ma
  • *Corresponding author for this work

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

Abstract

High dynamic range (HDR) imaging has seen a lot of progress in recent years, while an efficient way to capture and generate HDR video is still in need. In this paper, we present a method to generate HDR video from frame sequence of alternating exposures in a fast and concise fashion. It takes advantage of the recent advancement in deep learning to achieve superior efficiency compared to other state-of-art method. By training an end-to-end CNN model to estimate optical flow between frames of different exposures, we are able to achieve dense image registration of them. Using this as a base, we develop an efficient method to reconstruct the aligned LDR frames with different exposure and then merge them to produce the corresponding HDR frame. Our approach shows good performance and time efficiency while still maintain a relatively concise framework.

Original languageEnglish
Title of host publicationImage and Graphics - 9th International Conference, ICIG 2017, Revised Selected Papers
EditorsYao Zhao, Xiangwei Kong, David Taubman
PublisherSpringer Verlag
Pages70-81
Number of pages12
ISBN (Print)9783319715971
DOIs
StatePublished - 2017
Externally publishedYes
Event9th International Conference on Image and Graphics, ICIG 2017 - Shanghai, China
Duration: 13 Sep 201715 Sep 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10668 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference9th International Conference on Image and Graphics, ICIG 2017
Country/TerritoryChina
CityShanghai
Period13/09/1715/09/17

Keywords

  • Convolutional neural networks
  • High dynamic range video
  • Optical flow

Fingerprint

Dive into the research topics of 'Efficient high dynamic range video using multi-exposure CNN flow'. Together they form a unique fingerprint.

Cite this