Real-time surveillance video display with salience

Abstract

In this paper, we aim at providing a means for efficient display of surveillance video. In video surveillance, usually, there are certain regions of interest (ROIs), such as entrance or exit, and moving objects or persons, which should be paid more attention. By tracking and locally zooming the ROI, the proposed method adds salience on it to help surveillant to locate such important region effectively. Here, salience means highlighting special region locally. Given an input video signal, the ROI is detected first. The original video frame is mapped as a texture on a deformed mesh to produce the zoom effect. The position and shape of the ROI determine the mesh deformation. By experiment, we show that the proposed method is effective and efficient.