| 基于不变矩和支持向量机的手势识别 |
| |
| 引用本文： | 董立峰,阮军,马秋实,汪雷.基于不变矩和支持向量机的手势识别[J].微型机与应用,2012,31(6):32-35. |
| |
| 作者姓名： | 董立峰 阮军 马秋实 汪雷 |
| |
| 作者单位： | 武汉理工大学信息工程学院,湖北武汉,430070 |
| |
| 摘 要： | 选取Hu不变矩、手势轮廓的凹陷个数及其周长与面积比为手势识别的主要特征,采用了基于径向基核的SVM分类器进行0～9十种手势的识别。实验结果表明,在背景单一、光照情况良好条件下,该方法具有很高的识别率,并且简单快速。
|
| 关 键 词： | 手势识别 Hu矩 支持向量机 |
The gesture identification based on invariant moments and SVM |
| |
| Dong Lifeng,Ruan Jun,Ma Qiushi,Wang Lei.The gesture identification based on invariant moments and SVM[J].Microcomputer & its Applications,2012,31(6):32-35. |
| |
| Authors: | Dong Lifeng Ruan Jun Ma Qiushi Wang Lei |
| |
| Affiliation: | (College of Information Engineering,Wuhan University of Technology,Wuhan 430070,China) |
| |
| Abstract: | This paper selects Hu moment invariants,the sag number of gestures outline and the ratio of its perimeter and area as the main characteristics of gesture recognition,adopts SVM classifier based on the RBF kernel to recognize ten kinds gestures of 0~9.The result of the experiment shows that this method has a high recognition rate under a single background and good illumination condition,and it is more simple and faster. |
| |
| Keywords: | gesture identification Hu moments support vector machine |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
