电子学报 ›› 2020, Vol. 48 ›› Issue (1): 118-123.DOI: 10.3969/j.issn.0372-2112.2020.01.014

• 学术论文 • 上一篇    下一篇

基于显著性特征的选择性目标跟踪算法

丁新尧, 张鑫   

  1. 华南理工大学电子与信息学院, 广东广州 510640
  • 收稿日期:2018-11-07 修回日期:2019-05-06 出版日期:2020-01-25 发布日期:2020-01-25
  • 通讯作者: 张鑫
  • 作者简介:丁新尧 男,1996年生于河南南阳.本科毕业于华南理工大学,现为华南理工大学电子与信息学院硕士研究生,研究方向为计算机视觉与图像处理.E-mail:scutdxy@163.com
  • 基金资助:
    广东省科技发展专项基金(No.2016A010101014);广东省自然科学基金(No.2018A030313295)

Visual Tracking with Salient Features and Selective Mechanism

DING Xin-yao, ZHANG Xin   

  1. School of Electronics and Information Engineering, South China University of Technology, Guangzhou, Guangdong 510640, China
  • Received:2018-11-07 Revised:2019-05-06 Online:2020-01-25 Published:2020-01-25
  • Supported by:
     

摘要: 针对长期目标跟踪算法中目标部分遮挡甚至消失情况下的目标有效跟踪问题,提出了一种融合了目标显著性特征的选择性跟踪算法.首先,为了有效抑制背景信息的干扰,综合HOG特征以及颜色统计特征的特点提出了前景概率图来实现增强目标显著性抑制背景干扰的效果.其次,为了减少跟踪漂移和解决重度照明和遮挡等挑战性场景中的跟踪失败问题,引入了具有筛选条件的选择性跟踪和检测框架,用以控制检测器的激活以及最终结果的选择.OTB2013数据集上的实验结果证明,本文算法可以取得91.1%的总体准确率以及67%的总体成功率,结果优于大部分跟踪算法.

 

关键词: 前景概率图, 条件检测机制, 跟踪置信度, 特征显著性

Abstract: In the long time tracking,object representation and occlusion handling are two important challenges.We propose a selective tracking and detection framework in which a new probabilistic object-enhanced feature is integrated.Firstly,we propose a foreground probability map to enhance the target and weaken the surrounding background.Secondly,we introduce the selective tracking and detection framework that has two sets of conditions to control the detector activation and final result selection.We have evaluated our methods on the popular benchmark OTB2013 dataset.The algorithm achieves an overall accuracy of 91.1% and a success rate of 67%,which demonstrates that our algorithm performs favorably compared with other state-of-the-art methods.

Key words: foreground probability map, condition detection mechanism, tracking confidence, salient features

中图分类号: