Multi-View Video Summarization

Abstract

Video Camera use to capture videos and photos that may be compressed after that, saved and transmitted to a location with a limited number of monitors. For security and environmental monitoring, all public buildings such as banks, educational institutions, offices, and hospitals now have several cameras installed with the field of view overlapping. Video summarization is a method of generating a summary of entire video content using still photos or video skimming. The length of the summary video should be smaller than that of the original video and it should include all of the highest information frames from the original video. Video summaries are more difficult to create because the video recorded by cameras is longer, also, include non-informative events, and the same scene is recorded multiple times with different views because of that inter-view dependencies and illumination variations in this paper BIRCH clustering algorithm used for remove redundant and static frames. After that apply k means clustering algorithm for finding key frames and generate final video summary.