Spatio-temporal descriptor for abnormal human activity detection

Fam Boon Lung, Mohamed Hisham Jaward, Jussi Paavo Samuli Parkkinen

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

7 Citations (Scopus)

Abstract

There has been an increased interest in the field of abnormal human activity detection to find a good descriptor with a lower computational cost. In this paper, we propose such a Spatio-Temporal Descriptor (STD) based on spatio-temporal features of an image sequence. Proposed descriptor is based on a texture map, known as Spatio-Temporal Texture Map (STTM) and is based on 3-dimensional Harris function. It is able to capture subtle variations in the spatio-temporal domain. Performance of the STD was illustrated with a mixture of Gaussian Hidden Markov Model (HMM) to show its potential for more complex modeling. Proposed algorithm was evaluated with UCSD dataset that has abnormal events that are not staged such as biker, skater, cart activities etc. Compared to other state of the art descriptors that are used with the same dataset, our proposed descriptor shows competitive performance with a lower computational cost.
Original languageEnglish
Title of host publicationProceedings of the 14th IAPR International Conference on Machine Vision Applications, MVA 2015
EditorsNorimichi Ukita, Eigo Segawa, Norichika Yui
Place of PublicationNew Jersey USA
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages471 - 474
Number of pages4
ISBN (Print)9784901122146
DOIs
Publication statusPublished - 2015
EventMachine Vision Applications 2015 - Tokyo Japan, Tokyo, Japan
Duration: 18 May 201522 May 2015
Conference number: 14

Conference

ConferenceMachine Vision Applications 2015
Abbreviated titleMVA 2015
Country/TerritoryJapan
CityTokyo
Period18/05/1522/05/15
Other14th IAPR International Conference on Machine Vision Applications, MVA 2015

Cite this