华中科技大学学报（自然科学版）2011，Vol.39Issue(z2)：204-207,219,5.
基于判别式序列表的均值漂移目标跟踪算法
Mean shift object tracking algorithm based on discriminative ranked list
摘要
Abstract
As for the interference of the background information around the tracking object, an improved mean shift tracking algorithm was proposed based on discriminative ranked list, in which, the object template and candidate objects were represented using the object appearance feature and the discriminative ranked list was employed to model for the object appearance information and describe the background information around the object being tracked. Based on the framework of mean shift tracking algorithm, we integrated the object appearance and the discriminative ranked list object model to improve the traditional mean shift tracking algorithm by means of reducing the interference of the background information. Finally, the proposed tracking approach and the mean shift tracking algorithm were tested on several image sequences. The experimental results show that the proposed method can improve the mean shift tracking algorithm effectively and has higher tracking accuracy and more robust tracking efficiency.关键词
目标跟踪/均值漂移/判别式序列表/目标外观/背景信息Key words
object tracking/ mean shift/ discriminative ranked list/ object appearance/ background information分类
信息技术与安全科学引用本文复制引用
蒋良卫,桑农,刘乐元..基于判别式序列表的均值漂移目标跟踪算法[J].华中科技大学学报（自然科学版）,2011,39(z2):204-207,219,5.基金项目
国家自然科学基金重点资助项目(60736010) （60736010）
中央高校基本科研业务费资助项目(2010ZD034) （2010ZD034）
广东省教育部产学研结合项目(2009B09200047). （2009B09200047）
