Comparison of Scene Change Detection Algorithms for Videos

Abstract

A Video Scene Change Detection method is needed in order to manage the rising amount of video content efficiently in applications like digital library, object tracking, watermarking, video on demand (VOD), etc. Scene change detection is done by various methods like pixel-based, histogram, Otsu's, edge-based, statistics-based, compression-based, etc. In this paper, we present abrupt scene change detection for videos using pixel-based, edge-based and Otsu's method. The results are compared depending on the video size and time.