Automatic Appropriate Segment Extraction from Shots Based on Learning from Example Videos
Automatic Appropriate Segment Extraction from Shots Based on Learning from Example Videos
复制标题
DOI：
10.1007/978-3-540-92957-4_94
复制
发表时间：
2009-01
期刊：
影响因子：
--
通讯作者：
Yousuke Kurihara;Naoko Nitta;N. Babaguchi
中科院分区：
文献类型：
--
作者：
Yousuke Kurihara;Naoko Nitta;N. Babaguchi
视频是由镜头组成的，每个镜头都是由摄像机连续记录的，视频编辑可以看作是从原始视频中选择镜头重新排序的过程。镜头通常包括多余的间隔，这些间隔通常由专业编辑剪掉。本文提出了一种从镜头中自动提取合适片段的方法，将编辑视频中使用的完整间隔定义为合适片段，将其他长度相等的间隔定义为不合适片段。由于不同内容的镜头适合在编辑后的视频中使用何种特征的间隔是不同的，因此本文提出的方法首先使用支持向量机对镜头根据内容进行分类。然后，利用隐马尔可夫模型对每个镜头类别学习到的合适和不合适片段的音频和视觉特征的时间模式，提取合适的片段。通过实验验证了该方法的有效性。
Videos are composed of shots, each of which is recorded continuously by a camera, and video editing can be considered as a process of re-sequencing shots selected from original videos. Shots usually include redundant intervals, which are often edited out by professional editors. Defining the intact interval which is used in the edited video as the appropriate segment and all other intervals of equal length as inappropriate segments, this paper proposes a method for automatically extracting appropriate segments from shots. Since what kinds of characteristics make an interval appropriate to be used in the edited video should be different among shots with different content, the proposed method firstly categorizes shots according to their content with Support Vector Machines. Then, the appropriate segments are extracted based on the temporal patterns of audio and visual features in appropriate and inappropriate segments learned with Hidden Markov Models for each shot category. The effectiveness of the proposed method is verified with experiments.