A video coloring method based on CNN and feature point tracking

George Guan*, Fuquan Zhang, Gangyi Ding, Meng Niu, Lin Xu

*Corresponding author for this work

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

2 Citations (Scopus)

Abstract

Black and white films were the main form of human culture records before. Colorization of those films is creative. At present, Colorization of black and white films is still handmade which is expensive and time consuming. In this paper, a framework based on CNN and particle filter tracking algorithm is proposed, which can color black and white video and try to solve the problem of dynamic frame based on context correlation. At the same time, the objective function of CNN structure and particle filter tracking are optimized. The result of colorization on videos is satisfactory.

Original languageEnglish
Title of host publicationAdvances in Smart Vehicular Technology, Transportation, Communication and Applications - Proceedings of the 1st International Conference on Smart Vehicular Technology, Transportation, Communication and Applications
EditorsJeng-Shyang Pan, Tsu-Yang Wu, Yong Zhao, Lakhmi C. Jain
PublisherSpringer Science and Business Media Deutschland GmbH
Pages312-320
Number of pages9
ISBN (Print)9783319707297
DOIs
Publication statusPublished - 2018
Event1st International Conference on Smart Vehicular Technology, Transportation, Communication and Applications, VTCA 2017 - Kaohsiung, Taiwan, Province of China
Duration: 6 Nov 20178 Nov 2017

Publication series

NameSmart Innovation, Systems and Technologies
Volume86
ISSN (Print)2190-3018
ISSN (Electronic)2190-3026

Conference

Conference1st International Conference on Smart Vehicular Technology, Transportation, Communication and Applications, VTCA 2017
Country/TerritoryTaiwan, Province of China
CityKaohsiung
Period6/11/178/11/17

Keywords

  • Black and white films
  • CNN structure
  • Colorization
  • Optimized
  • Particle filter tracking

Fingerprint

Dive into the research topics of 'A video coloring method based on CNN and feature point tracking'. Together they form a unique fingerprint.

Cite this