Published Paper Details:

The popularity of digital video is increasing rapidly. To help users navigate in huge data base, the algorithms that automatically index video based on key frames are needed. One approach is to extract key frames based on clustering approach. Key frame is a frame which can represent the salient content of the shot. Key frames provide a suitable abstraction and framework for video indexing, browsing and retrieval. They allow users to quickly browse over the video by viewing only a few high-lighted frames. The use of key-frames greatly reduces the amount of data required in video indexing and provides an organizational framework for dealing with video content. Because of its importance, much research effort has been given in key frame extraction. The presented key frame extraction scheme relies on an improved spectral clustering algorithm for the clustering of frame sequences of video. Video frame are clustered into groups by their similarity differences. Then the clustering process is carried for shots of the video using spectral clustering method. Then the key frame identification scheme is established for key frame extraction process and key frame of the video shot are extracted. Finally proposed model presents a key frame extraction method based on clustering scheme using the spectral clustering technique and the key frames of the shot are extracted.