Abstract:

This paper presents a video surveillance system that displays mega-pixel streams effectively, while traansmitting and processing the streams efficiently with limited resources such as bandwidth, computing power and display resolution. The proposed system stores high-resolution and high-quality video data and associated object metadata, which includes ROI (Region-of-Interest)information. To satisfy such resource constraints and display important parts in detail without missing the overall scene context, the stored images are efficiently transcoded in the compressed-domain based on the ROI information, display resolution and available bandwidth. Simulation results confirm the effectiveness of the proposed system in terms of objective measures and subjective evaluation.