A new video editing interface based on the automatic extraction of emphasized speech and 'laughter' scenes

摘要

Most current video skimming engines create a video digest automatically with fixed and closed algorithms and do not really support authoring; users understand and control the skimming process only imperfectly. This paper proposes a new video editing interface, called Digest Editor (DE), that integrates two skimming engines based on the automatic extraction of emphasized speech and laughter scenes; it allows the setting of digest parameters and the simply manual editing of the resulting digests. Our goal is to realize a more convenient, flexible and effective video digest system for non-professionals. We perform an experiment to assess DE usability. The results show that users can understand and use our interface easily and effectively.