Cosyne 2008 Workshops
March 3-4, 2008
Snow Bird, Utah
Speaker Name
T. Serrre
Talk Title
A Biologically Inspired System for Action Recognition
Talk Abstract
We present a biologically-motivated system for the recognition of actions from video sequences. The approach builds on recent work on object recognition based on hierarchical feedforward architectures (Serre et al, 2005; Mutch & Lowe, 2006; Ranzato et al, 2007) and extends a neurobiological model of motion processing in the visual cortex (Giese & Poggio, 2003). The system consists in a hierarchy of spatio-temporal feature detectors of increasing complexity: An input sequence is first analyzed by an array of motion-direction sensitive units which, through a hierarchy of processing stages, lead to position and scale invariant space-time template feature detectors. We experiment with different types of motion-direction sensitive units as well as different system architectures. As in (Mutch & Lowe, 2006), we find that sparse features in intermediate stages outperform dense ones and that using a simple feature selection approach leads to an efficient system that performs better with far less features. We test the approach on different publicly available action datasets, in all cases achieving the highest results reported to date.
