Key-frame extraction and video summarization using QR-Decomposition

This paper explores the problem of high quality key frame extraction and video summarization detection and examines a novel video summarization algorithm by using QR-decomposition. We derive some efficient measures to compute the dynamicity of video shots using QR-Decomposition and we utilize it in detecting the number of keyframes could be selected from each shot. Also, we conclude a corollary that illustrates a new property of QR-Decomposition. We utilize this property in order to summarize video shots with low redundancy. The proposed algorithm is implemented and evaluated on TRECVID 2006 benchmark platform. The results confirm the high performance of the proposed algorithm.