A Robust and Efficient Framework for Video Shot Boundary Detection

Abstract

Due to the increase in the development of different technologies used in the video analysis, a huge number of videos file are presented for users on the internet. Hence, complexity increases in finding the video with desired contents. Shot Boundary Detection (SBD) is a very first step and important one in video browsing, indexing and retrieval. Different video shots are detected by extracting a feature s from individual video frames. Different Shot Boundary Detection Algorithms have been developed, but the major issues related to this are, they are inefficient and has high computational cost. In this paper, we proposed fast video shot boundary detection technique which utilizes the pre-processing scheme before shot detection which involves Candidate Segment Selection (CSS) and Bisection Based Comparison (BSC). Due to this an excessive non-boundary segments i.e., which doesn't carry any shot boundaries are eliminated and only candidate segments are further processed for actual which speed-up the shot detection.