Comparison of Algorithms for Cut and Fade Transition Detection for Video

Abstract

Mutual information is common information and similarity metric between two consecutive frames. In this work mutual information is also used as similarity metric of video frames. Mutual information is used mainly in cut detection; its gradual transitions detection performance is so well so, dynamic window-based mutual information shot boundary detection method is applied. This work presents a cut and fade transitions detection method based on mutual information and joint entropy. The experiments based on the proposed method show it reduces effectively shot error detection. Also we use mutual information and joint entropy to detect fade in/out in experiments. Cumulative mutual information use for detect dissolve effect from video.