Output details
13 - Electrical and Electronic Engineering, Metallurgy and Materials
Queen Mary University of London : A - Electrical and Electronic engineering
A dynamic texture-based approach to recognition of facial actions and their temporal models.
Highly cited work is part of a long-term collaboration with Prof. Pantic (Imperial College). It is the first work on facial action unit recognition using dynamic textures with clear benefits in recognising subtle activations of facial muscles. This was a central output of the EU Network of Excellence "Peer-to-Peer Tagged Media" (FP7-216444, 2008-2011) which supported Koelstra and of ERC Starting grant ERC-2007-StG-203143 supporting Pantic. Among others, it is referenced in the seminal paper (google citations=153) describing the extension of the most used dataset on facial expressions (Cohn-Kanade dataset, CVPR 2011, 10.1109/CVPRW.2010.5543262).