基于LSSVM的静态手势识别  

Static gesture recognition based on LSSVM

作　　者：段洪伟[1] 陈一民[1] 林锋[1] 

机构地区：[1]上海大学 计算机工程学院,上海200072

出　　处：《计算机工程与设计》2004年第12期2352-2353,2368,共3页Computer Engineering and Design

摘　　要：支持向量机(Support Vector Machine,简称SVM),是基于统计学习理论的一种新的模式识别方法,较好地解决了小样本学习问题。通过使非线性空间变换为线性空间,降低了算法的复杂性。LSSVM(Least Squares Support Vector Ma-chine)由于使用线性等式代替了标准的SVM算法中的线性不等式,进一步降低了运算量。利用傅立叶描述子获取静态手势特征向量,通过LSSVM大尺度算法求解方程组来得到LSSVM分类器,进行静态手势识别,取得了较高的识别率。说明如何把静态手势识别结果应用到机器人远程控制中,提高人机交互的友好性。Support vector machine (abbreviated as SVM) is a new method for pattern recognition based on statistical learning theory, which can solve the small-sample learning problem better. In addition, this theory convert the problem in non-linearity space to that in linearity space by using the kernel function, which reduces the complexity of the algorithm. LSSVM (least squares vector machine), reduces the complexity further through replacing the inequality by equality. The FFT descriptor is used to get the feature vector, and then the LSSVM as the classifier through the large scale algorithm to solve the equation set is used to the static gesture recognition, and get a good recognition rate. It is also explained on how to apply the static gesture recognition to the robot remote control for improving the natural interaction between man and robot.

关 键 词：手势识别 傅立叶描述子 分类器 统计学习理论 算法 机器人 支持向量机 线性空间 特征向量 求解 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP181[自动化与计算机技术—计算机科学与技术]