Key Frames Extraction Using Spline Curve Fitting for Online Video Summarization

Video summarization methods produce a video abstraction that permits users to obtain an informative video frames with minimum storage space and in less time. The keyframes is significantly characterize the salient contents of the video. This paper presents a design of video summarization framework for Internet videos to provide a quick way to realize, browse and review its contents. The main objective is to determine, extract and collect the most informative frames in the acquired video to formulate a summary video related to the original video. We have suggested a suitable approach for shot boundary detection along video frames. The sudden change benchmark between successive frames has calculated based on spline curve representation of frame data points inspired by capturing the visual difference. As well as, a selection and integration of the key frames from each shot is specified through clustering the higher differences between shot frames, in order to tackle the redundant frames issue and generate the video summary. From the experimental results, the proposed video summary approach is capable of capturing an informative content of video shots and preventing redundancy frames with minimum requirements in terms of storage space.