Who, what, when, where, why and how in video analysis: an application centric view

This paper presents an end-user application centric view of surveillance video analysis and describes a flexible, extensible and modular approach to video content extraction. Various detection and extraction components including tracking of moving objects, detection of text, faces, and face based soft biometric for gender, age and ethnicity classification are described within the general framework for real-time and post event analysis applications Panoptes and VideoRecall. Some end-user applications that are built on this framework are discussed.