摘要
在指纹录入以及预处理中易引入噪声,因此在特征提取中,存在大量的伪特征,影响指纹的特征信息提取。本文提出了一种新的伪特征去除算法,即利用特征点为中心建立特征点检测窗口,根据一定的条件对伪特征点进行删除,另外本文提出了一种利用计算特征点到中心点的纹线数取代计算这两者之间的距离的特征提取算法,更加精确地记录了指纹的实际特征,为提高指纹识别的精度奠定了良好的基础。
Because of the of noises in the fingerprint input and pretreatment, there are a lot of pseudo-features in fingerprint minutiae extraction .This paper proposes a new algorithm for removing pseudo-features, namely using minutiae point as a center to create a minutiae point examination window and delete pseudo-features according to certain circumstances. The paper also proposes an improved algorithm by computing the fingerprint lines between feature point and the central point instead of computing the distance of this two points.
出处
《自动化技术与应用》
2008年第7期33-35,共3页
Techniques of Automation and Applications
关键词
指纹
细节特征提取
伪特征
特征向量
fingerprints
minutiae extraction
pseudo-features
feature vector
