Adaptive Threshold based Video Shot Boundary Detection Framework

Shot boundary detection (SBD) is the first step in a video management system. Threshold determination and gradual transition detection are two challenging problems that this paper will focus on. To address the first problem, we investigate an adaptive threshold determination technique. It builds a statistical model according to the video frame differences and relates each threshold adaptively to the specified video itself. For the second problem, we propose a dissolve type determination algorithm, which uses adaptive two-threshold with transparency and edge computation to detect a dissolve transition. Then, we further propose an adaptive threshold based shot boundary detection framework using wavelet transform for frame thumbnails extraction and between cluster distances to smooth noises. Extensive experiments have been carried out on the platform of TRECVID benchmark. The experimental results show that the proposed technique outmatches conventional methods in F-cut by 0.4% and in F-gradual by 17.45%.