This paper presents a smart surveillance system named CASSANDRA, aimed at detecting instances of aggressive human behavior in public environments. A distinguishing aspect of CASSANDRA is the exploitation of complementary audio and video cues to disambiguate scene activity in real-life environments. From the video side, the system uses overlapping cameras to track persons in 3D and to extract features regarding the limb motion relative to the torso. From the audio side, it classifies instances of speech, screaming, singing, and kicking-object. The audio and video cues are fused with contextual cues (interaction, auxiliary objects); a Dynamic Bayesian Network (DBN) produces an estimate of the ambient aggression level. Our prototype system is validated on a realistic set of scenarios performed by professional actors at an actual train station to ensure a realistic audio and video noise setting.

@en

This article presents an overview of current automatic sign recognition research. A review of recent studies, as well as on our own research, has identified several problem areas that hamper successful sign recognition by a computer. Some of these problems are shared with automatic speech recognition, whereas others seem to be unique to automatic sign recognition. These latter difficulties include context dependency, determination of the basic units of modeling, the ability to distinguish between signs and gestures, movement epenthesis, and repetition within signs. As a possible solution to these problems, bottom-up processing should perhaps be supplemented with top-down processing.@en