Robust human behavior modeling from multiple cameras

Kosmopoulos, DI; Voulodimos, AS; Varvarigou, TA

dc.contributor.author	Kosmopoulos, DI	en
dc.contributor.author	Voulodimos, AS	en
dc.contributor.author	Varvarigou, TA	en
dc.date.accessioned	2014-03-01T02:46:58Z
dc.date.available	2014-03-01T02:46:58Z
dc.date.issued	2010	en
dc.identifier.issn	10514651	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/32960
dc.subject	Behavior modeling	en
dc.subject	Fusion	en
dc.subject	Hidden Markov models	en
dc.subject	Recognition	en
dc.subject.other	Behavior modeling	en
dc.subject.other	Fusion	en
dc.subject.other	Gaussians	en
dc.subject.other	Heavy occlusion	en
dc.subject.other	Human behavior modeling	en
dc.subject.other	Human behaviors	en
dc.subject.other	Illumination changes	en
dc.subject.other	Motion history images	en
dc.subject.other	Multiple cameras	en
dc.subject.other	Multivariate Student	en
dc.subject.other	Real environments	en
dc.subject.other	Recognition	en
dc.subject.other	Target recognition	en
dc.subject.other	Training algorithms	en
dc.subject.other	Training data	en
dc.subject.other	Visual behavior	en
dc.subject.other	Cameras	en
dc.subject.other	Feature extraction	en
dc.subject.other	Hidden Markov models	en
dc.subject.other	Tracking (position)	en
dc.subject.other	Video cameras	en
dc.subject.other	Behavioral research	en
dc.title	Robust human behavior modeling from multiple cameras	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1109/ICPR.2010.872	en
heal.identifier.secondary	http://dx.doi.org/10.1109/ICPR.2010.872	en
heal.identifier.secondary	5597830	en
heal.publicationDate	2010	en
heal.abstract	In this work, we propose a framework for classifying structured human behavior in complex real environments, where problems such as frequent illumination changes and heavy occlusions are expected. Since target recognition and tracking can be very challenging, we bypass these problems by employing an approach similar to Motion History Images for feature extraction. Furthermore, to tackle outliers residing within the training data, which might affect severely the training algorithm of models with Gaussian observation likelihoods, we scrutinize the effectiveness of the multivariate Student-t distribution as the observation likelihood of the employed Hidden Markov Models. Additionally, the problem of visibility and occlusions is addressed by providing various extensions of the framework for multiple cameras, both at the feature and at the state level. Finally, we evaluate the performance of the examined approaches under real-life visual behavior understanding scenarios and we compare and discuss the obtained results. © 2010 IEEE.	en
heal.journalName	Proceedings - International Conference on Pattern Recognition	en
dc.identifier.doi	10.1109/ICPR.2010.872	en
dc.identifier.spage	3575	en
dc.identifier.epage	3578	en