Creating memorable video summaries that satisfy the user's intention for taking the videos
Abstract: Highlights•We proposed memorability-based video segmentation, a method that is efficient and accurate for unedited user videos.•We introduce three scores to generate dynamic video summaries: image memorability score, snap point score, and motion score. The generated dynamic summaries are not only interesting but also capture the user’s intention for taking the video.•Experimental results indicate that our proposed video summarization method produces video summaries that have higher importance scores and better satisfy the viewers’ preference than the state-of-the-art methods.
Loading