摘要： 为挖掘视频中丰富的语义信息，提出基于负样本精简概念格规则的语义概念检测方法。分析基于概念格的语义分析系统，考虑训练数据中负样本的信息，提出利用负样本精简的语义规则提取算法，将其应用于视频语义检测。先将视频镜头的低层特征映射到低层语义特征，再利用该算法生成语义分类规则，进行视频语义概念检测。实验结果表明，该方法是有效可行的。

关键词: 概念格, 视频语义概念检测, 精简规则, 低层语义特征, k-均值

Abstract: To mine the rich semantic information in videos, this paper proposes a semantic concept detection method based on concept lattice rules reduced by negative samples. Through analyzing the semantic analysis system based on concept lattice and considering the information of negative samples in training data, it designs an algorithm for extracting semantic rules reduced by negative samples and apples the algorithm in video semantic detection. The low-level features of the video shots are mapped to the low-level semantic features, and the video semantic concept is detected after the semantic classification rules are generated by the proposed method. Experimental results show the feasibility and validity of the method.

Key words: concept lattice, video semantic concept detection, reduced rule, low-level semantic feature, k-means

中图分类号: 

• TP311