一种基于低分辨率指纹图像的增强方法

指纹鉴别是一个比较复杂的过程,一般包括指纹采集、图像预处理、特征提取及特征匹配等几个步骤。其中图像预处理中的图像增强是最为关键的环节,直接影响着后面指纹图像的细化特征提取等处理。论文针对硅晶体指纹采集芯片采集到的低分辨率指纹图像,提出一种有效的图像增强处理方法。由于指纹分辨率低,以及前景与背景的灰度差别很大,采用通常的空域Gabor增强后仍然有误增强和增强不充分的情况。所以针对此情况论文提出了采用非线性与线性相结合的二次图像增强的方法。首先使用Gabor滤波器进行非线性的图像增强,对增强后的灰度图像进行二次线性增强,即利用指纹的方向信息进行二值化处理。实验结果表明论文方法增强效果较好,与经典的Gabor图像增强方法比较,使用该方法图像增强后特征提取准确率明显提高。

An Enhancement Algorithm Based on Low Resolution Fingerprint Image

Fingerprint verification is a complex process,which includes fingerprint image capture,image pretreatment,minutia extraction,minutia matching,etc.A critical step is the fingerprint enhancement that affects the process of thinning and minutia extraction directly.This paper presents a fingerprint enhancement algorithm,which can improve the clarity of image captured by low resolution sensor.Because there is some noise in a fingerprint image,Gabor filters usually don't remove the noise completely or enhance the image by mistake.The algorithm contains not only linear enhancement algorithm but also nonlinear enhancement algorithm.First the fingerprint image will be enhanced nonlinearly by Gabor filter.For the enhanced image,this paper improves the clarity of image by linear enhancement,that is two value processes based on the estimated local ridge orientation.Experimental results show that the enhancement algorithm improves accuracy of minutia extraction.