Skip to main navigation Skip to search Skip to main content

Dynamic region selection in video based on spatio-temporal multiple instance learning

  • Xiao Zheng Wang
  • , Xu Dong Zhao
  • , Peng Liu
  • , Xiang Long Tang

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

Abstract

The selection of dynamic region in video plays an important role in many subsequent vision-based applications, especially in scene classification with different weather conditions. In this paper, we extract five local features from pixel blocks of each frame in a video, and propose an approach to dynamic region selection based on a presented description of spatio-temporal multiple instances. The effectiveness of our method is shown using experiments on videos under different weather environments.

Original languageEnglish
Title of host publication23rd International Conference on Computer Graphics and Vision, GraphiCon 2013 - Conference Proceedings
PublisherGraphiCon Scientific Society
Pages103-105
Number of pages3
ISBN (Electronic)9785804414024
StatePublished - 2020
Externally publishedYes
Event23rd International Conference on Computer Graphics and Vision, GraphiCon 2013 - Vladivostok, Russian Federation
Duration: 16 Sep 201320 Sep 2013

Publication series

Name23rd International Conference on Computer Graphics and Vision, GraphiCon 2013 - Conference Proceedings

Conference

Conference23rd International Conference on Computer Graphics and Vision, GraphiCon 2013
Country/TerritoryRussian Federation
CityVladivostok
Period16/09/1320/09/13

Keywords

  • Dynamic region
  • K-means
  • Multi-instance
  • Spatio-temporal feature

Fingerprint

Dive into the research topics of 'Dynamic region selection in video based on spatio-temporal multiple instance learning'. Together they form a unique fingerprint.

Cite this