Andrew Zisserman - Human Focussed Video Analysis

Опубликовано 18 августа 2016, 4:54

Determining the pose and actions of humans is one of the central problems of image and video analysis. The visual problem is challenging because humans are articulated animals, wear loose and varying clothing, self-occlude themselves, and stand against difficult and confusing backgrounds. Nevertheless, the area has seen great progress over the last decade due to advances in modelling, learning, and in the efficiency of algorithms. We describe approaches for recognizing human actions and interactions, and for determining 2-D upper body pose. Results will be shown for various TV videos and feature films, and a live demonstration given for pose based video retrieval. This is joint work with Vitto Ferrari, Nataraj Jammalamadaka, C. V. Jawahar, Alexander Klaeser, Marcin Marszalek, Alonso Patron-Perez, Ian Reid, and Cordelia Schmid.