自适应的基于Gabor函数的指纹图像增强算法  

Adaptive Fingerprint Enhancement Algorithm Based on Gabor Filter

作　　者：苑玮琦[1] 刘兆悦[1] 

机构地区：[1]沈阳工业大学视觉检测研究所,辽宁沈阳110023

出　　处：《激光与红外》2008年第8期844-848,共5页Laser & Infrared

基　　金：国家自然科学基金项目(No.60472088)资助

摘　　要：针对实际应用时采集的指纹库中指纹图像纹线粗细和质量差别很大的情况,提出了一种根据纹线宽度自适应地决定Gabor滤波器变换窗大小的指纹图像增强算法。改进了方向图的算法,在小波变换之后的低频子图上进行方向估算。实验结果表明,基于低频于图计算方向图使得运算速度和结果准确性均有提高,尤其对于低质量的指纹图像因为小波变换滤除噪声的作用,方向图准确性提高更为明显;自适应决定变换窗大小使Gabor滤波器能更好地发挥连接断线和分离粘连的作用,处理过粗或过细指纹时更具优越性。In fingerprint database collected in daily life there are ridges and valleys whose widths differ much . An a- daptive Gabor based fingerprint enhancement algorithm was proposed to adapt images with different texture sizes. The methods for ridge orientation are improved by the processing on the subimage. Experimental results indicate that calcu- lating orientation image by subimage has the speed and verasity advantages which are obvious on low quality images for the noise-reducing function of wavelet transfermation. Adaptive Gabor filter works better on thinner or thicker fin- gerprint textures for its better joint of faulted ridges and separation of aeeretionary ridges.

关 键 词：GABOR滤波器 指纹图像增强 方向图 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]