KIT at MediaEval 2015 - Evaluating Visual Cues for Affective Impact of Movies Task
P Marin Vlastelica, Sergey Hayrapetyan, Makarand Tapaswi, Rainer Stiefelhagen · MediaEval · 2015
We present the approach and results of our system on the MediaEval Aective Impact of Movies Task. The challenge involves two primary tasks: aect classication and violence detection. We test the performance of multiple visual features followed by linear SVM classiers. Inspired by successes in dierent vision elds, we use (i) GIST features used in scene modeling, (ii) features extracted from a deep convolutional neural network trained on object recognition, and (iii) improved dense trajectory features encoded using Fisher vectors commonly used in action recognition.