Multi-view Video Summarization Algorithm for WMSN

摘要

Wireless Multimedia Sensor Networks (WMSN) has been one challenge category of special wireless sensor networks. Different-view videos captured by cameras in various different locations are typical multi-view videos. In order to reduce transmission quantity and conserve nodes' energy, traditional multi-view video coding approaches are long suitable for WMSN. In this paper, we propose a multi-view video summarization algorithm for WMSN, where different-view videos are converted to one monocular video with consideration of spatio and temporal correlations of networks, so as to reduce the redundancy of multi-view video. Moreover, different-view videos of adjacent sensor nodes are coded with correlated arrangement of I frames, B frames and P frames. Finally, a few preliminary simulations are present.