摘要： 镜头检测算法的研究是视频分析与检索中较为活跃的研究领域之一。该文提出了一种基于运动信息的镜头切变检测算法。该算法采用四叉树算法分析视频序列的运动信息，进而提取基于运动信息的特征矢量，由提取的特征矢量可以检测出视频序列中的镜头切变点。实验表明该算法是有效的，可行的，并且具有非常强的抗干扰能力。

关键词: 镜头检测, 运动信息, 四叉树算法

Abstract: The study on shot transition detection is an active research field in video analysis and retrieval. A novel algorithm based on motion information is suggested to apply for shot cut detection. This method firstly adopts quad-tree algorithm to analyze motion information involving in the video sequences, and hence abstracts the feature vector concerning on motion information. The shot cut detection can be achieved for the specified video sequences by using the feature vector. It is verified by the experimental results that this proposed approach is feasible and efficient for shot cut detection, especially with fortified capability of noise immunity.

中图分类号: 

• TP391.41