摘要
提出一种基于轮廓波(Contourlet)变换与支持向量机(SVM)的掌纹识别算法。基于积分光密度与中心矩,对掌纹图像进行光照、位置与方向的归一化,提取Contourlet变换高频子带的一阶统计特征,形成掌纹特征,利用SVM进行分类与识别。实验结果表明,与基于统计特征的掌纹识别方法相比,该算法的识别率较高。
This paper describes palmprint verification based on the Contourlet transform and Support Vector Machine（SVM）. Palmprint images are normalized in the orientation, position and illumination conditions based on the integrated optical density, central moments, and one order statistics of each sub-band are calculated in their Contourlet domains and regarded as features. A SVM-based classifier is employed to implement recognition. Experimental results show that the recognition rate is higher than that of palmprint recognition algorithms based on statistical features.
出处
《计算机工程》
CAS
CSCD
2012年第6期196-197,200,共3页
Computer Engineering
基金
陕西省自然科学基础研究计划基金资助项目(2009JM8003)
关键词
CONTOURLET变换
支持向量机
掌纹识别
统计特征
中心矩
Contourlet transform
Support Vector Machine（SVM）
palmprint recognition
statistical feature
central moment
