• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2014, Vol. 36 ›› Issue (08): 1549-1554.

• 论文 • 上一篇    下一篇

基于视觉注意机制的目标跟踪方法

张巧荣,张慧娟   

  1. (河南财经政法大学计算机与信息工程学院,河南 郑州 450002)
  • 收稿日期:2012-10-16 修回日期:2013-03-26 出版日期:2014-08-25 发布日期:2014-08-25
  • 基金资助:

    河南省基础与前沿技术研究计划资助项目(122300410379);河南省教育厅科学技术研究重点资助项目(14A520025)

Object tracking based on visual attention           

ZHANG Qiaorong,ZHANG Huijuan   

  1. (College of Computer and Information Engineering,Henan University of Economics and Law,Zhengzhou 450002,China)
  • Received:2012-10-16 Revised:2013-03-26 Online:2014-08-25 Published:2014-08-25

摘要:

提出一种基于视觉注意机制的运动目标跟踪方法。该方法借鉴人类的视觉注意机制的研究成果,建立视觉注意机制的计算模型,计算视频中各部分内容的视觉显著性。结合视觉显著性计算结果,提取视频图像中的显著性目标。利用颜色分布模型作为目标的特征表示模型,与视频中各显著目标进行特征匹配,实现目标的跟踪。在多个视频序列中进行实验,并给出相应的实验结果及分析。实验结果表明,提出的目标检测与跟踪算法是正确有效的。

关键词: 目标检测, 目标跟踪, 视觉注意, 视觉显著性, 显著图

Abstract:

An algorithm for moving object tracking based on visual attention mechanism is proposed. Human visual attention mechanism is introduced in the proposed algorithm and a computational model of visual attention is constructed. Based on the computational model, the visual saliency of each part in videos is computed. Salient objects are detected according to the visual saliency results. Using the color distribution model as the representation model of the object, objects are tracked by calculating the similarity of salient objects in the two successive frames by. The algorithm is tested on many video sequences. Experimental and analysis results show that the proposal is robust and effective for object detection and tracking in videos.

Key words: Object detection;object tracking;visual attention;visual saliency;saliency map