| 低阶不变矩人耳识别方法 |
| |
| 引用本文： | 王晓云,郭金玉.低阶不变矩人耳识别方法[J].沈阳大学学报,2012,24(2):62-65. |
| |
| 作者姓名： | 王晓云 郭金玉 |
| |
| 作者单位： | 1. 沈阳理工大学机械学院,辽宁沈阳,110168
2. 沈阳化工大学信息工程学院,辽宁沈阳,110142 |
| |
| 基金项目： | 辽宁省博士启动基金资助项目 |
| |
| 摘 要： | 介绍了不变矩原理及特点,由于高阶矩对噪声敏感,提出一种低阶不变矩分子区域人耳识别方法.对分子区域的人耳图像,提取各个区的低阶不变矩首尾相连组成一组特征矢量,作为耳识别模型.在北京科技大学建立的图像库遍历实验后,结果表明,低阶矩识别效果好于高阶矩,分区好于整体.划分32区低阶不变矩达到100%的识别率.
|
| 关 键 词： | 灰度归一化 低阶矩 子区域 耳识别 |
Human Ear Recognition Methods based on Low-order Moment Invariants |
| |
| WANG Xiaoyun,GUO Jinyu.Human Ear Recognition Methods based on Low-order Moment Invariants[J].Journal of Shenyang University,2012,24(2):62-65. |
| |
| Authors: | WANG Xiaoyun GUO Jinyu |
| |
| Institution: | 1. School of Mechanical Engineering, Shenyang Ligong University, Shenyang 110168, China 2. School of Information Engineering, Shenyang University of Chemical Technology, Shenyang 110142, China) |
| |
| Abstract: | The principle and characteristics of moment invariants are introduced. Because higher-order moment invariants are sensitive to noise, a new recognition method is proposed based on low-order moment invariants at sub-region of human ear images. Ear feature vector is composed of low-order moment invariants of each sub-region. Each ear image is divided into different sub-region, the result shows that it is the most effective ear recognition method by ear image divided 32 sub-region, and it achieves 100% recognition rate on USTB ear database. |
| |
| Keywords: | gray normalization low-order moment invariants sub-region ear recognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
