一种基于解压的镜头探测方法

A Decompression-Based Shot Detection Method

镜头探测是视频内容和结构化分析的第一步。目前的镜头探测方法根据处理角度不同 ,主要分为压缩域和解压域两种。提出了一种基于解压的镜头探测方法。该方法基于这样一个假定 ,即同一镜头往往具有相同的背景区域。通过对背景区域的数据进行分析 ,得到视频的镜头边界。该方法具有设置参数少、阈值敏感度低以及处理数据量小的特点 ,适用于大容量的视频库。通过对各种不同类型的视频进行实验 ,证明该方法具有较强的适应性和鲁棒性 。

Shot detection is the first step in video content structured analysis. According to different processing purposes, the existing shot detection methods are mainly based on compression domain and decompression domain. In this paper, a shot detection method based on decompression is presented. The method is based on the assumption that the same shots always have similar background area. Thus through analyses of the background area, shot boundaries are detectable. This method needs to set only a few parameters, is not sensitive to the predefined values, and processes much fewer data. Experiments on various types of videos indicate that the proposed method is both effective and robust, and can achieve high detection and precision ratios.