University of Cambridge > > Microsoft Research Machine Learning and Perception Seminars > Topic Models for Human Activity Understanding

Topic Models for Human Activity Understanding

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Microsoft Research Cambridge Talks Admins.

This event may be recorded and made available internally or externally via Microsoft will own the copyright of any recordings made. If you do not wish to have your image/voice recorded please consider this before attending

Automated modelling of activity in visual surveillance and unstructured consumer multi-media data are important capabilities for security and commercial content based indexing. These tasks are challenging because tracking and segmentation may be unreliable in with crowds and occlusion; manual labelling of sufficient training data may be prohibitively expensive; and interesting behaviours may be difficult to model due to being visually subtle relative to background activity and/or defined by complex interactions between multiple objects evolving over time. In this talk, I will describe our work addressing activity modelling with probabilistic topic models in both weakly supervised and unsupervised contexts. Within this framework, we can address unsupervised or weakly supervised anomaly detection, hierarchical activity mining and classification, generalising from sparse examples and active learning for rare activity discovery.

This talk is part of the Microsoft Research Machine Learning and Perception Seminars series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


© 2006-2022, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity