Development of Soft Segmentation Algorithm for Video Summarization

Abstract

newline Shot boundary detection (SBD), also known as temporal video segmentation, is newlineone of the techniques for Video summarization which refers to abstract the main newlineoccurrences, scenes, or objects in a clip in order to provide an easily interpreted newlinesynopsis Shot boundary detection (SBD) is the process of identifying the newlinetransitions between the adjacent shots. With the rapid growing amount of newlinemultimedia, content-based information retrieval has become more important. newlineFeatures including the pixel difference, color histogram, motion and edge newlineinformation etc., have been widely used to detect shot cuts inside videos. Despite newlineactive existing research, it still remains a challenge to achieve accurate SBD with newlineone single algorithm. newlineAutomatic shot boundary detection is always been a challenge for the newlineresearchers due to the variety of videos. There are so many varieties in video newlinedatabase like, different frame rates, different illumination, added computer newlineartifacts, different digitization techniques, different genres i.e, movies, sports, newlineanimation etc. it has been very difficult to have a universal technique to newlineeffectively segment the video and automatically detect the video shot boundaries newlineusing conventional techniques. Soft computing approaches are proven effective newlinefor carry out such task. newlineSoft computing approach is proposed which integrates hybrid features for SBD newlineinside general videos as well as to detect the shot boundaries. The approach is newlinebased on the concept to threshold the histogram according to the similarity newlinebetween gray levels assessed through a fuzzy measure. The advantage of newlineproposed techniques aims to overcome limitation of most of the conventional newlinemethods. Proposed soft computing based model, synthesizes unsupervised fuzzy newlinecompetitive learning algorithm and self-organized competitive network for newlinedetecting abrupt as well as gradual video shot boundary. newlineThe proposed algorithm uses the inherent characteristics of Soft computing newlinetechniques to automatically detect the video shot boundaries. The C