FAR at MediaEval 2013 Violent Scenes Detection: Concept-based Violent Scenes Detection in Movies
Mats Sjöberg, Jan Schlüter, Bogdan Emanuel Ionescu, Markus Schedl · MediaEval · 2013
The MediaEval 2013 Aect Task challenged participants to automatically nd violent scenes in a set of popular movies. We propose to rst predict a set of mid-level concepts from low-level visual and auditory features, then fuse the concept predictions and features to detect violent content. We deliberately restrict ourselves to simple general-purpose features with limited temporal context and a generic neural network classier. The system used this year is largely based on the one successfully employed by our group in the 2012 task, with some improvements based on our experience from last year. Our best-performing run with regard to the ocial metric received a MAP@100 of 49.6%.