Design, analysis, and engineering of video monitoring systems: An approachand a case study

Rapid improvement in computing power cheap sensing, and more flexible algor ithms are facilitating increased development of real-time video surveillanc e and monitoring systems. The deployment of video understanding systems in certain critical applications in the real world can be done only if perform ance guarantees can be provided for these systems. This paper reviews past work on a systematic engineering methodology for vision systems performance characterization and illustrates how it can be adapted in practice to deve lop a real-time people detection and zooming system to meet given applicati on requirements. A case study involving dual-camera real-time video surveil lance is used to illustrate that by judiciously choosing the system modules and by performing a careful analysis of the influence of various tuning pa rameters on the system it is possible to perform proper statistical inferen ce, to automatically set control parameters and to quantify performance lim its.