Dynamic Hierarchical Dirichlet Process for abnormal behaviour detection in video

Olga Isupova, Danil Kuzin, Lyudmila Mihaylova

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (SciVal)
17 Downloads (Pure)

Abstract

This paper proposes a novel dynamic Hierarchical Dirichlet Process topic model that considers the dependence between successive observations. Conventional posterior inference algorithms for this kind of models require processing of the whole data through several passes. It is computationally intractable for massive or sequential data. We design the batch and online inference algorithms, based on the Gibbs sampling, for the proposed model. It allows to process sequential data, incrementally updating the model by a new observation. The model is applied to abnormal behaviour detection in video sequences. A new abnormality measure is proposed for decision making. The proposed method is compared with the method based on the non-dynamic Hierarchical Dirichlet Process, for which we also derive the online Gibbs sampler and the abnormality measure. The results with synthetic and real data show that the consideration of the dynamics in a topic model improves the classification performance for abnormal behaviour detection.

Original languageEnglish
Title of host publication2016 19th International Conference on Information Fusion (FUSION)
PublisherIEEE
Number of pages8
ISBN (Electronic)978-0-9964-5274-8
ISBN (Print)978-1-5090-2012-6
Publication statusPublished - 4 Aug 2016
Event19th International Conference on Information Fusion, FUSION 2016 - Heidelberg, Germany
Duration: 5 Jul 20168 Jul 2016

Conference

Conference19th International Conference on Information Fusion, FUSION 2016
Country/TerritoryGermany
CityHeidelberg
Period5/07/168/07/16

Fingerprint

Dive into the research topics of 'Dynamic Hierarchical Dirichlet Process for abnormal behaviour detection in video'. Together they form a unique fingerprint.

Cite this