Skip to main navigation Skip to search Skip to main content

Evolvement Constrained Adversarial Learning for Video Style Transfer

  • SUNY Albany
  • JD Finance AI Lab
  • General Electric

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

4 Scopus citations

Abstract

Video style transfer is a useful component for applications such as augmented reality, non-photorealistic rendering, and interactive games. Many existing methods use optical flow to preserve the temporal smoothness of the synthesized video. However, the estimation of optical flow is sensitive to occlusions and rapid motions. Thus, in this work, we introduce a novel evolve-sync loss computed by evolvements to replace optical flow. Using this evolve-sync loss, we build an adversarial learning framework, termed as Video Style Transfer Generative Adversarial Network (VST-GAN), which improves upon the MGAN method for image style transfer for more efficient video style transfer. We perform extensive experimental evaluations of our method and show quantitative and qualitative improvements over the state-of-the-art methods.

Original languageEnglish
Title of host publicationComputer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers
EditorsKonrad Schindler, Greg Mori, Hongdong Li, C.V. Jawahar
PublisherSpringer Verlag
Pages232-248
Number of pages17
ISBN (Print)9783030208868
DOIs
StatePublished - 2019
Event14th Asian Conference on Computer Vision, ACCV 2018 - Perth, Australia
Duration: Dec 2 2018Dec 6 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11361 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference14th Asian Conference on Computer Vision, ACCV 2018
Country/TerritoryAustralia
CityPerth
Period12/2/1812/6/18

Fingerprint

Dive into the research topics of 'Evolvement Constrained Adversarial Learning for Video Style Transfer'. Together they form a unique fingerprint.

Cite this