摘要
综述了当前基于Kinect手势跟踪的几种代表性方法:基于轮廓和特征匹配运动目标跟踪算法、均值漂移跟踪算法、连续自适应均值漂移跟踪算法、卡尔曼滤波跟踪算法以及粒子滤波跟踪算法,分析其特点及相互关系,详细探讨了各方法的优缺点。最后对手势跟踪发展趋势作出了展望。
The paper reviewed tile current representative algorithms of hand gesture tracking based on Kinect, namely moving larget tracking algorithm based on contour and feature matching, Mean Shift tracking algorithm, continuous adaptive Mean Shift tracking algorithm, Kalman filter tracking and particle filter tracking algorithm, and analyzed their characteristics. Then it discussed the advantages and disadvantages of these algorithms carefully. Finally, this paper prnposed the prnspect on the development trends of hand gesture tracking.
出处
《计算机应用研究》
CSCD
北大核心
2015年第7期1921-1925,共5页
Application Research of Computers
基金
国家自然科学基金资助项目(61203316
61203319)
江苏省自然科学基金资助项目(BK20141002)
虚拟现实技术与系统国家重点实验室开放基金资助项目(BUAA-VR-13KF-04)
关键词
KINECT
人机交互
手势跟踪
匹配搜索
粒子采样
Kineet
human-computer interaction
tracking of hand gesture
match search
particle sampler
