基于隐含马尔可夫模型网络的视频识别方法

Video Clip Recognition Using Hidden Markov Model Networks

提出用隐含马尔可夫模型网络(Hidden Markov Model Networks)描述和识别短视频序列(如广告视频)。实验结果表明,该方法不仅对转码产生的图像质量鲁棒性好,而且可以很好识别截短的视频序列。

This paper proposes using Hidden Markov Model （HMM） Networks to describe and recognize short video clips, e.g. advertisement clips. Results show that the gained image has better robustness, and the method can also recognize short version of a video clip very well.