Modular approach for multiple event detection in surveillance videos

Imran Ahmed, Ghulam Masood, Qazi Nida-Ur-rehman, Muharnmad Nawaz, Zahoor Elahi

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

1 Citation (Scopus)

Abstract

In the prevailing law and order situation video surveillance have widespread applications from public places to monitoring and security. In this paper a modular approach has been proposed to detect multiple events in videos. We have divided these events into three broad categories i.e. intrusion, loitering and slip and fall. The proposed approach is divided into primary and secondary analysis. Videos for surveillance have to be passed through the primary analysis, which can be used as an input for the secondary analysis. The proposed system achieved higher accuracy (90 %) than state of the art (85 %) for the aforementioned events. The results have been verified using publically available benchmark datasets.

Original languageEnglish
Title of host publicationWCE 2016 - World Congress on Engineering 2016
EditorsLen Gelman, David W.L. Hukins, S. I. Ao, S. I. Ao, Len Gelman, S. I. Ao, Alexander M. Korsunsky, Andrew Hunter
PublisherNewswood Limited
Pages484-489
Number of pages6
ISBN (Electronic)9789881925305
Publication statusPublished - 2016
Externally publishedYes
EventWorld Congress on Engineering 2016, WCE 2016 - London, United Kingdom
Duration: 29 Jun 20161 Jul 2016

Publication series

NameLecture Notes in Engineering and Computer Science
Volume2223
ISSN (Print)2078-0958

Conference

ConferenceWorld Congress on Engineering 2016, WCE 2016
Country/TerritoryUnited Kingdom
CityLondon
Period29/06/161/07/16

Keywords

  • Abnormal events
  • Linear kaiman filter
  • Video surveillance

Fingerprint

Dive into the research topics of 'Modular approach for multiple event detection in surveillance videos'. Together they form a unique fingerprint.

Cite this