Shot Detection Using Pixel wise Difference with Adaptive Threshold and Color Histogram Method in Compressed and Uncompressed Video

Abstract

To elaborate a video in terms of its content, it needs to be partitioned into its smallest visual unit called shot. To segment a video into shots, shot boundary is needed. Here pixel wise difference is adapted for uncompressed and compressed video shot transition detection with adaptive threshold. Color histogram is also used for the cut and dissolve boundary. For cut boundary, traditional threshold technique works well but it fails in some cases like camera flash, fast zooming etc. To improve its performance 2nd derivative method is accepted for cut detection and for the dissolve boundary histogram difference with twin comparison method is used. Pixel intensity based approach is utilized for fade detection.