基于几何矩的自适应手势跟踪算法研究  

Research of self-adaption gesture tracking algorithm based on the geometric moment

作　　者：周荣强[1] 全海燕[1] 熊志伟[1] 张天飞[1] 

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650093

出　　处：《郑州轻工业学院学报（自然科学版）》2013年第3期51-55,共5页Journal of Zhengzhou University of Light Industry:Natural Science

基　　金：云南省自然科学基金项目(2009ZC048M);昆明理工大学校人才培养基金项目(KKZ3201103022)

摘　　要：在Mean-Shift算法基础上,提出了一种新算法:用肤色检测与寻找最大连通域相结合的方法来自动选取跟踪的手势;采用图像几何矩替代Mean-Shift矢量,可以有效地跟踪到手势目标而且可以实现跟踪窗尺度随手势目标大小、角度及形状的改变而改变.实验结果表明,本算法成功地对动态且有形变的手势进行了实时跟踪,并且捕捉到其运动轨迹.算法平均每s约处理27帧图像,可以满足实时性要求.A new algorithm was proposed which based on the Mean-Shift algorithm. This algorithm can automatically select tracking gestures by combining skin color detection with seeking maximum connected do- main;it using image geometric moment to alternate the Mean-Shift vector, and can also effectively track gesture target and achieve the size of tracking window changing along with the gesture＇s change of size, angle and shape. The experimental results showed that deformable gestures and capture its trajectory in real second, and can meet the real-time requirements. the algorithm can successfully track the dynamic and time. This algorithm processes about 27 frames persecond, and can meet the real-time requirements.

关 键 词：手势跟踪 MEAN-SHIFT算法 连通域 肤色检测 几何矩 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]