A detection-based person tracking algorithm

被引：0
作者：
机构：
[1] College of Mechatronics Engineering and Automation, National University of Defense Technology
来源：
Hu, D. (dwhu@nudt.edu.cn) | 1600年 / National University of Defense Technology卷 / 36期
关键词：
Agglomerative clustering; KLT tracker; Object tracking; Person detection; Real-time;
D O I：
10.11887/j.cn.201402019
中图分类号：
学科分类号：
摘要：
The traditional object tracking algorithms require manually annotated tracking area, and suffer from the problem of drift. To address these difficulties, the problem of person tracking was focused on, and a new detection-based tracking algorithm was proposed. To reduce failure in tracking, multiple detectors to locate multiple body parts were employed, and then their detection results were mapped to a common body area. To adapt for the quickly moving objects, the KLT tracker and agglomerative clustering for linking the detection windows to form person body trajectories was employed. The experimental results reveal that using multiple detectors improves the tracking performance significantly, and the KLT tracker is adaptable for quickly moving objects. This algorithm is nearly real-time.
引用
收藏
页码：113 / 117
页数：4
相关论文
共 15 条
  • [11] Benenson R., Mathias M., Timofte R., Et al., Pedestrian detection at 100 frames per second, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2903-2910, (2012)
  • [12] Patron-Perez A., Marszalek M., Zisserman A., Et al., High five: Recognizing human interactions in tv show, Proceedings of the British Machine Vision Conference, Aberystwyth, (2010)
  • [13] Everingham M., Gool L., Williams C.K., Et al., The Pascal visual object classes challenge 2007 (VOC) results, International Journal of Computer Vision, 88, 2, pp. 303-338, (2010)
  • [14] Prest A., Schmid C., Ferrari V., Weakly supervised learning of interactions between humans and objects, IEEE Transactions on Pattern Analysis and Machine Intelligence, 34, 3, pp. 601-614, (2012)
  • [15] Eichner M., Ferrari V., Better appearance models for pictorial structures, Proceedings of the British Machine Vision Conference, (2009)
12