以影片內容為基準之視訊摘要系統

成崑佑
The paper presents a new framework of video summarization which can extract and summarize video shots that a user interests in from a long and complicated video, according to their similarity of motion type and scene. Firstly, the shot detection adopts the color and edge information to make shot boundaries accurate.

Then the clustering process classifies the shots according to their similarities in scenes and motion types. Finally, we select the important shots of each cluster by estimating their priority value. The priority value determines the importance of each shot by measuring the motion energy and color variation. The proposed method can produce a classified video summary, which allows users to review and search the video more easily. Experiment results illustrate that the proposed method can successfully classify a video into several clusters of different motion types and scenes, and extract the specific shots according to their importance.

以影片內容為基準之視訊摘要系統論文摘要：在這篇論文中，我們提出了一個新的影片簡介系統架構，這個系統可以針對使用者所輸入的影片來選取他們感興趣的精采片段並做為整段影片的簡介。在本系統中會先將影片作整體內容的分析以及場景變換的偵測藉以把影片分割成許多段不同的鏡頭(shot)，再選取每一個鏡頭中的關鍵畫面(key-frame)作為鏡頭分類(shot clustering)的依據。在分類的時候是根據每一個關鍵畫面的場景(scene)以及動作(motion type)作為根據，如果兩兩的鏡頭擁有相似的場景以及動作，那麼它們將會被分類到同一群(cluster)裡面。使用我們所提出的分類方法之後，使用者可以得到許多由不同場景以及動作所構成的群。之後，本系統會根據每一個鏡頭的顏色變化程度(color variation)以及物體的動作大小(motion energy)選出每一個群裡面的精采畫面(highlight)來重新建構這段影片的簡介(summary)，並根據精彩畫面的重要性來決定這些鏡頭在最後的簡介中可以播放的長度，這麼一來可以讓使用者所得到的影片簡介更加的簡潔。