In this working note, we propose a set of features and a classification scheme for detecting automatically violent scenes in movies. The features are extracted from audio, video, and subtitles modalities of the movies. In violent scenes classification, we found the following features relevant: the short time audio energy, motion component, and shot words rate.We classified the shots into violent and non-violent using naive Bayesian, Linear Discriminant Analysis (LDA), and Quadratic Discriminant Analysis (QDA) targeting to maximize the precision of the detection in the first two minutes of retrieved content.