Study Of Video Stabilization And Shot Boundary Detection Algorithm Based On Feature Windows Tracking

With the rapid development of multimedia and Internet, the more and morevideo information can be obtained easily. How to organize, manage and index theseinformation leads to a new research field of video processing, content-based videoretrieval and indexing. The first important task of content-based video retrieval andindexing is the shot boundary detection. With the video acquisition devices increaseddramatically, video stabilization, the removal of unwanted vibrations in a videosequence induced by undesirable carrier motion, is an essential part of videoacquisition in industrial, military and consumer applications. The emphasis of thispaper is put on the above two topics.They look like disrelated, but there is a same problem contained in these twofields, namely motion parameter calibration of camera. Vibration of images due tocamera shake is stabilized with motion compensation using motion parameter. Andto distinguish the differences of image content changing due to the shots transitioneffect and camera motion, which is the essential problem in shot boundary detection.So feature windows tracking algorithm is systematically investigated firstly in thispaper, then the proposed algorithm extracts the corner-points as features by SUSANalgorithm and tracking these feature windows by fuzzy Kalman filtering. Afterresearching and analyzing the principle and some techniques of video stabilizationand shot boundary detection, the feature windows tracking algorithm is adopted withcombining some others measures and approaches.The experimental results illustrate that the proposed image stabilizationalgorithm is effective and robust with low computational complexity. And theproposed shot boundary detection algorithm has the ability both to detect shotboundary and to identify and locate the boundary. The proposed algorithms haveimportant applied and research values.