A Neuro-Fuzzy Technique for Scene Cut Detection from MPEG-1 Video

摘要

In this paper we propose a Neuro-Fuzzy (NF) algorithm that can automatically detect the type of MPEG frame at which clear scene cuts occurred. The algorithm is based on detection of the majority of Intra-macroblocks and the directionality of the two neighboring B-frames motion vector types from the bit stream. It is shown that the method is able to identify the location of scene cuts in I- P-and each one of B-frames with a very high success rate.