A Novel Dynamic Video Summarization Approach Based on Rough Sets in Compressed Domain

摘要

In this study, a novel dynamic video summarization approach based on Rough Sets (RS) is developed. It can not only rapidly generate video summary that minimizes the visual content redundancy for input video sequences, but also specify the number of frames to get various summary according to user demand. First, DCT coefficients and DC coefficients, the most important video visual features are extracted from raw video sequences to represent video information. Second, an Information system is constructed with DC coefficients. Third, a new and concise Information system is achieved by using the reduction theory of RS, meanwhile, the effective representation of frames and its corresponding reduced frame numbers are recorded, i.e., dynamic video summarization. Experimental results indicate that the proposed algorithm is more effective and intelligent than conventional methods in video summarization generation.