摘要
指纹识别是生物特征识别技术中的热点。指纹特征可分为全局特征与细节特征,现有主流指纹识别方法是基于细节特征的识别,但是指纹全局特征识别具有明显的优势,更加符合人类识别机理。本文尝试采用PCA网络提取指纹的全局特征——主分量特征,理论分析与实验说明了指纹主分量特征是有效的,鲁棒性较好。在识别方法上,采用协同模式识别方法,该方法注重模式的整体信息,并将其与PCA网络特征提取层有机的结合,对FVC2002指纹库的实验表明,本文的指纹全局特征识别方法预处理与特征提取简单,识别速度快,鲁棒性较好,取得了良好的指纹识别与身份认证效果。
Fingerprint recognition is one of the research hotspots of biometrics techniques. Generally, there are two kinds of features in person's fingerprints: global features and minutiae features. And the leading fingerprint recognition methods are based on minutiae features of fingerprints. But fingerprint global feature recognition has some obvious advantages, and it is similar to human recognition mechanism. PCA neural network is adopted to extract global features of fingerprint images in this paper. We analyze and testify that fingerprint principal components features are effective and robust. A novel and effective synergetic pattern recognition approach is adopted in the recognition layer, which emphasizes global information of patterns and combines with PCA neural network very well. With FVC2002 fingerprint database, experiment results show that the purposed method is simple, fast and robust.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2004年第1期87-93,共7页
Pattern Recognition and Artificial Intelligence
基金
国家自然科学基金(No.60175011
60375011)
安徽省自然科学基金(No.01042301)
安徽省优秀青年科技基金(No.04042044)
关键词
生物特征识别
指纹识别
PCA网络
指纹特征
指纹图像
图像分割
Fingerprint Recognition, Synergetic Pattern Recognition, PCA Neural Network, Feature Extraction, Global Feature
