Informative Content Extraction through Key Frames

Abstract

Nowadays, a huge amount of multimedia data is available on internet. As there is much redundancy, we can’t go through all this data. For various purposes this data is browsed, retrieved and processed. But when time and speed constraints are taken into consideration, accessing the data is very inefficient. Video Summarization is one of the way which makes it possible to give non-redundant, effective, feature based abstract view of the video which is covering entire contents in terms of the selected key frames. Feature based selected frames will lead to the summarized video. There are two main technique of video summarization i.e. key frame based and video skimming. This paper focuses on key frame extraction using video abstraction, visual descriptors and bag of visual words approach.