An Algorithm for Shot Boundary Detection and Key Frame Extraction Using Histogram Difference

Shot boundary detection is the complete segmentation of a video into continuously imaged temporal video segments. Condensed video representation is the extraction of video frames or short clips that are semantically representative of the corresponding video. Both tasks are very significant for the organization of video data into more manageable forms. In order to extract valid information from video, process video data efficiently, and reduce the transfer stress of network, more and more attention is being paid to the video processing technology. The amount of data in video processing is significantly reduced by using video segmentation and key-frame extraction. So, these two technologies have gradually become the focus of research. With the square histogram difference considered at block level for the video frames, a new method of extracting the keyframes based on shot type is presented. Keywords—Shot Boundary Detection, KeyFrame Extraction.