Our first presenter of the new semester is our very own Ehsan Amjadian.
Of the many diverse topics his research covers, Ehsan will talk to us about “Syntactic Pattern Recognition in Dynamic Data”. Please see below the abstract of the talk.
Abstract— It has been shown that syntactic patterns are useful in classification of visual data. For instance it has been shown that Context Free Trees can be used to preserve hierarchical structures in visual data, and HMM’s are beneficial in recording the linear order in dynamic visual data. Yet this talk introduces another way that syntactic patterns can be exploited to help the recognition task. The present work uses Stochastic Linear Formal Grammar to add real-time post-estimation capabilities to the pattern recognition task in a computationally efficient way.