没有合适的资源?快使用搜索试试~ 我知道了~
首页区域协方差跟踪的混合搜索策略提升:高效与鲁棒性
区域协方差跟踪的混合搜索策略提升:高效与鲁棒性
0 下载量 57 浏览量
更新于2024-08-26
收藏 338KB PDF 举报
本文主要探讨了区域协方差跟踪技术在计算机视觉领域的最新进展,特别是在提高效率和鲁棒性方面所面临的挑战。传统的协方差跟踪依赖于基于区域协方差的特征描述符,这种方法在目标识别和跟踪任务中表现出色,但由于其蛮力搜索策略(brute-force search)的低效性,可能导致跟踪轨迹不连续,对噪声和干扰敏感。 为了克服这些问题,作者提出了一种混合搜索策略,该策略结合了多种优化技术。首先,整数区域计算被引入,这种方法比传统的积分图像更快,因为它能够自适应地处理不同大小和形状的目标,同时适应跟踪环境的变化。其次,搜索过程采用了从粗到细的策略,这样可以快速排除大部分无效区域,减少计算量。 圆形搜索是另一种加速手段,它减少了搜索空间,使得搜索更加聚焦于可能的目标位置。此外,均值漂移优化也被整合进来,这是一种基于概率模型的迭代方法,有助于在局部区域内找到最有可能的目标位置,从而提高跟踪的稳定性和精度。 在论文中,作者强调了通过在正常稳态下进行局部搜索,他们的方法不仅提高了效率,而且显著增强了对目标表示的多样性,使得算法具有更好的鲁棒性。实验证明,这种方法在处理特定的视频序列时,不仅能够实现更快速的跟踪,还展现出更强的抗干扰能力。 这篇研究论文旨在革新协方差跟踪技术,通过混合搜索策略的创新设计,提升其在实际应用中的性能,如自动驾驶、无人机监控、视频监控等领域。这项工作的贡献在于提供了一个更为高效、稳定和鲁棒的区域协方差跟踪框架,为后续的研究者提供了新的思路和技术手段。
资源详情
资源推荐
Int. J. Embedded Systems, Vol. 7, No. 1, 2015 55
Copyright © 2015 Inderscience Enterprises Ltd.
Region covariance tracking with hybrid
search strategy
Ruhan He*
Institute for Pattern Recognition and Artificial Intelligence,
Huazhong University of Science and Technology,
Wuhan 430074, China
Email: heruhan@hotmail.com
*Corresponding author
Yongsheng Yu
The Green Building Materials and Manufacturing Engineering
Research Center of Ministry of Education,
Wuhan University of Technology,
Wuhan 430070, China
Email: yongshengyu@whutedu.cn
Jia Chen, Min Li and Xun Yao
Collaborative Innovation Center of Common Technology
in Textile Industry Chain in Hubei Province,
Wuhan Textile University,
Wuhan 430073, China
Email: chenjiawh@sina.com
Email: lindali@wtu.edu.cn
Email: 810408384@qq.com
Abstract: Covariance tracking has achieved impressive successes owing to its competent region
covariance-based feature descriptor. However, the brute-force search strategy in covariance
tracking is inefficient and it possibly leads to inconsecutive tracking trajectory and distraction. In
this work, a hybrid search strategy for optimisation on covariance tracking is proposed. The
hybrid strategy contains the integral region computation, the coarse-to-fine and circular search,
and mean shift optimisation. The integral region is much faster than integral image and adaptive
to the tracking target and tracking condition. The other three dramatically speed up the
convergence of the search. The proposed approach yields much better efficiency, robustness of
distraction, and stable trajectory by local search in normal steady state. Our approach shows
excellent target representation ability, faster speed, and more robustness, which has been verified
on some video sequences.
Keywords: covariance tracking; integral region; coarse-to-fine; CTF; circular search; mean shift;
hybrid search strategy.
Reference to this paper should be made as follows: He, R., Yu, Y., Chen, J., Li, M. and Yao, X.
(2015) ‘Region covariance tracking with hybrid search strategy’, Int. J. Embedded Systems,
Vol. 7, No. 1, pp.55–62.
Biographical notes: Ruhan He is an Associate Professor in College of Mathematics and
Computer Science, Wuhan Textile University, Wuhan, China. He is also a Post Doctor at the
Institute for Pattern Recognition and Artificial Intelligence, Huazhong University of Science and
Technology,Wuhan, China.. He received his MSc and PhD from Huazhong University of Science
and Technology (HUST) in 2004 and 2007, respectively. His research focuses on machine
learning, multimedia processing and analysis, etc. He has authored more than 30 scientific papers
and articles. He also initiated and led over multiple research projects.
Yongsheng Yu is currently a researcher in The Green Building Materials and Manufacturing
Engineering Research Center of Ministry of Education at Wuhan University of Technology. He
received his BS in Computer Science from Wuhan University of Hydraulic and Electrical
Engineering in 2001. He received his MS and PhD in Computer Science from Huazhong
University of Science and Technology in 2005 and 2010, respectively. His research interests
include computer security, image processing and computer simulation.
下载后可阅读完整内容,剩余7页未读,立即下载
weixin_38714509
- 粉丝: 3
- 资源: 931
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
最新资源
- 十种常见电感线圈电感量计算公式详解
- 军用车辆:CAN总线的集成与优势
- CAN总线在汽车智能换档系统中的作用与实现
- CAN总线数据超载问题及解决策略
- 汽车车身系统CAN总线设计与应用
- SAP企业需求深度剖析:财务会计与供应链的关键流程与改进策略
- CAN总线在发动机电控系统中的通信设计实践
- Spring与iBATIS整合:快速开发与比较分析
- CAN总线驱动的整车管理系统硬件设计详解
- CAN总线通讯智能节点设计与实现
- DSP实现电动汽车CAN总线通讯技术
- CAN协议网关设计:自动位速率检测与互连
- Xcode免证书调试iPad程序开发指南
- 分布式数据库查询优化算法探讨
- Win7安装VC++6.0完全指南:解决兼容性与Office冲突
- MFC实现学生信息管理系统:登录与数据库操作
资源上传下载、课程学习等过程中有任何疑问或建议,欢迎提出宝贵意见哦~我们会及时处理!
点击此处反馈
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功