摘要
介绍了基于CMOS图像传感器的便携式指纹采集装置的设计,利用全反射原理,并通过对指纹图像全局的方向分布获取特征向量。实验证明,特征向量融合方法使分类的准确率达到93.7%。
The design of a portable fingerprint image collecting device is introduced based on CMOS image sensor. In this device, the theory of total reflection is used, and the eigenvector are obtained by the general directional distributing of fingerprint images. The experimental results show that the accuracy of classification is improved to 93.7% by using the fused methods of feature extractions.
出处
《电视技术》
北大核心
2008年第2期84-86,共3页
Video Engineering
关键词
光电指纹图像采集
全局方向特征
图像滤波
中心点
opto-electronic fingerprint image collect
general directional feature
image filter
central points
作者简介
宋榕，硕士生，主研图像处理、模式识别；
林喜荣，教授,研究方向为基于生物特征的身份识别。
