Robust tracking via hough forests based on weak classifier response | |
Chen, Weiyi ; Dai, Pingyang ; Liu, Wei ; Li, Cuihua ; Xie, Yi ; Dai PY(戴平阳) ; Li CH(李翠华) | |
2013 | |
关键词 | Data processing Experiments Pixels Tracking (position) |
英文摘要 | Conference Name:14th Pacific-Rim Conference on Multimedia, PCM 2013. Conference Address: Nanjing, China. Time:December 13, 2013 - December 16, 2013.; Discriminative tracking method based on Hough forests has received increasingly attention due to its advantages such as its efficiency and robustness. But Hough forests only use the pixel values in the split test selection, the information of image patches has not been made full use of to discriminate different classes. In this paper, we propose a novel method of Hough forests for visual tracking. This method adopt a new strategy of the split test selection which is based on the weak classifier responses instead of pixel values. Compared with the traditional method of Hough forests, the proposed method can extract more information. Furthermore, this method introduces the LBP feature into Hough forests and then captures both edge-based shape structure and texture information. The proposed method is validated through the experiments on a set of publicly available benchmark dataset. The experiment results have shown that this method outperforms the state-of-the-art algorithms in terms of robustness and effectiveness. ? Springer International Publishing Switzerland 2013. |
语种 | 英语 |
出处 | http://dx.doi.org/10.1007/978-3-319-03731-8_26 |
出版者 | Springer Verlag |
内容类型 | 其他 |
源URL | [http://dspace.xmu.edu.cn/handle/2288/86572] |
专题 | 信息技术-会议论文 |
推荐引用方式 GB/T 7714 | Chen, Weiyi,Dai, Pingyang,Liu, Wei,et al. Robust tracking via hough forests based on weak classifier response. 2013-01-01. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论