VideoSense

	Kompetenzzentrum: Information Retrieval and Machine Learning
	Kontakt: Prof. Dr.-Ing. Sahin Albayrak, Esra Acar
	Website: http://www.videosense.eu/

The VideoSense "Network of Excellence" brings together leading European research groups for a deeper cooperation in video content analysis by considering both the technological and the ethical aspects of the data analytics approaches within security. In particular, VideoSense conducts joint collaborative research studies focused on the most challenging problems and outstanding issues to be resolved in the multi-modal content analysis of surveillance videos. With respect to these issues, one of the related research areas is event detection in video sequences.

In VideoSense, the DAI-Labor in collaboration with the other participating research groups aims at building a framework for multi-modal event detection in videos. The main goal is to analyze surveillance videos to extract valuable information. Specifically, this includes:

Analysis of features to determine the discriminative ones for video description,
Detection of pre-defined events by multi-modal analysis,
Detection of unusual events

The visual content of video data is described and classified by utilizing discriminating features (e.g. local or global feature descriptors) and groups of such features. In addition, audio data streams are analyzed using time series methods.

For further information, please visit http://www.videosense.eu/.