A hybrid scheme for temporal video segmentation

Abstract

Video temporal segmentation is normally the first and important step for content-based video applications. Although existing research on shot cut detection is active and extensive, it still remains a challenge to achieve accurate detection of all types of shot boundaries. In this paper, we propose a hybrid scheme to combine some features and techniques for detecting all sorts of shot cuts inside general videos. The hybrid scheme contains two processing modes, which are unified by a mode-selector to decide which mode the scheme should work on in order to achieve accurate temporal video segmentation. Via using the publicly test data set from Carleton University, we report extensive experimental results to evaluate the proposed algorithm. In comparison with existing algorithms, the experimental results support that the proposed algorithm outperforms the benchmark in terms of the precision-recall rates in detecting all types of shot cuts.