Cut detection using local image descriptor

Great advancement and popularity of multimedia technology facilitates uploading and accessing thousands of videos through internet in seconds. This demands adopting efficient video data management techniques for ease of storage and browsing applications. This in turn requires an efficient method of annotation. Automatic cut detection is the elementary and essential step of any automatic annotation process. The efficacy of intensity based cut detection and histogram based methods decrease with variation of intensity of the scene. In this paper we propose a novel cut detection approach based on the distance measure derived from differential excitation feature. The proposed distance measure has a strong spatial correlation among neighbouring pixels which is insensitive to the light variation. The proposed algorithm is tested with six test videos and compared with few existing popular approaches for validation.