基于模糊粗糙集的新闻视频镜头边界检测方法

镜头边界检测是实现基于内容的视频检索的一个重要步骤.为了将视频分割成镜头,现有的方法大都是首先提取大量的特征然后构造相异性测度函数.然而,太多的特征就会降低算法的效率.因此,有必要对镜头边界检测的规则进行特征约简.本文将粗糙集中的属性重要性和模糊粗糙集中的分类精度相结合定义了模糊粗糙算子,并构造了相异度检测函数.最后给出了镜头边界检测的一般性规则.由于本文检测方案的自适应性,因此适合于各种类型的新闻视频.用来自中央电视台的3个多小时的新闻视频所做的镜头边界检测实验获得了95.4%的查全率和96.1%的准确率.

A Shot Boundary Detection Method for News Video Based on Rough-Fuzzy Sets

As a crucial step in content-based news video indexing and retrieval system, shot boundary detection attracts much more research interests in recent years. To partition news video into shots, many metrics were constructed to measure the similarity among video frames based on all the available video features. However, too many features will reduce the efficiency of the shot boundary detection. Therefore, it is necessary to perform feature reduction for the decision rules of shot boundary. For this purpose, the classification method based on rough-fuzzy sets for feature reduction and the dissimilarity function is proposed. Finally, for each given class, the fuzzy if-then rule is generated for decision with fuzzy inference. The efficacy of the proposed method is extensively tested on more than 3 hour of CCTV news programs and 95.4% recall with 96.1% precision have been achieved.