• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2014, Vol. 36 ›› Issue (05): 941-946.

• 论文 • 上一篇    下一篇

基于手轮廓的深度图像手势识别方法

许凯,王敏   

  1. (河海大学计算机与信息学院,江苏 南京 211100)
  • 收稿日期:2012-11-30 修回日期:2013-01-21 出版日期:2014-05-25 发布日期:2014-05-25
  • 基金资助:

    教育部新世纪优秀人才支持计划(NCET100327);中央高校基本科研业务费项目(2009B21014);国家自然科学基金资助项目(30700183);教育部博士点基金项目(20070294001)

Hand gesture recognition in a depth
image based on hand contour        

XU Kai,WANG Min   

  1. (College of Computer and Information,Hohai University,Nanjing 211100,China)
  • Received:2012-11-30 Revised:2013-01-21 Online:2014-05-25 Published:2014-05-25

摘要:

提出了一种新的手势识别方法,该方法从深度图像中提取手形轮廓,通过计算手形轮廓与轮廓形心点的距离,使用离散傅里叶变换获得手势的表观特征,引入径向基核的支持向量机识别手势。建立了一个常见的10种手势的数据集,测试获得了97.9%的识别率。

关键词: 深度图像, 手势识别, 支持向量机, 离散傅里叶变换

Abstract:

A new hand gesture recognition method is proposed,which extracts hand contour from a depth image.The distance between hand contour and the center of contour is calculated, the representation feature of the hand gesture is obtained by using Discrete Fourier Transform (DFT),and the hand gesture is identified by using Support Vector Machine (SVM) with Radial Basis Function (RBF).A dataset with 10 common hand gestures are established in order to demonstrate the effectiveness of the proposed method and achieve about 97.9% recognition rate.
    

Key words: depth image;hand gesture recognition;support vector machine;discrete Fourier transform (DFT)