Video Skimming for Quick Browsing based on Audio and Image Characterization

摘要: Digital video is rapidly becoming an important source for information, entertainment and a host of multimedia applications. With the size these collections growing to thousands hours, technology needed effectively browse segments in short time without losing content video. We propose method extract significant audio information create “skim” which represents synopsis original. The extraction such as specific objects, keywords relevant structure, made possible through integration techniques image language understanding. resulting skim much smaller, retains essential original segment. This research sponsored by National Science Foundation under grant no. IRI9411299, Space Aeronautics Administration, Advanced Research Projects Agency. views conclusions contained this document are those authors should not be interpreted necessarily representing official policies or endorsements, either expressed implied, United States Government.