Coding efficiency of video sensor networks

We address the problem of coding multiple image sequences in a video sensor network. Multiple correlated video signals originate from cameras that monitor the same scene from different view points. Based on motion-compensated temporal transform coding of image sequences, the paper develops a high-rate approximation for the coding efficiency of a video sensor network with N cameras. We establish the rate difference between collaborative coding of correlated video signals and non-collaborative coding at each sensor. Bounds are obtained by assuming very accurate disparity compensation among all cameras. The goal is to study the impact of both inter-view correlation among video sensor signals and number of cameras in the network.