一种利用人物相似度的视频索引算法  

A Novel Video Indexing Approach Based on People-Similarity

作　　者：王鹏[1] 马宇飞[2] 张宏江[2] 杨士强[1] 

机构地区：[1]清华大学计算机科学与技术系,北京100084 [2]微软亚洲研究院,北京100080

出　　处：《电子学报》2004年第6期968-972,共5页Acta Electronica Sinica

摘　　要：本文面向视频分析和索引技术 ,提出一种利用人物相似度进行视频索引的算法 .该算法应用SVMs概率输出理论 ,将底层特征空间距离映射为语义层人物相似度 ,并提出一种新的非监督聚类算法 ,修正部分误判 ,最终实现对视频节目中人物的自动聚类和索引 .实验结果表明 ,该算法实用而高效 。This paper presents a novel approach to video indexing.In this approach,we define a people-similarity measure according to both clothing similarity and speaking voice similarity.Such similarity depicts how perceptually similar two people appearing in different scenes are and whether they belong to an identical person.The extended support vector machines are applied to map a serial of low-level feature distances to a perceived people-similarity.To build people-similarity based video indexing,a novel unsupervised clustering algorithm is also proposed,which can more correctly group individual person according to the mutual people-similarities among multiple people pairs.

关 键 词：视频分析和索引 人物相似度 机器学习 非监督聚类 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]