基于二值指纹图像的Gabor滤波增强算法

Gabor filter enhancement algorithm based on binarization fingerprint image

在指纹预处理过程中,Gabor滤波器对于图像的平滑和分割具有十分显著的效果。但是Gabor滤波器对纹线方向和频率十分敏感,通过处理灰度图像得到的二值图像存在着严重的特征点丢失和变化问题。采用相位差分二值化方法,使二值化过程不依赖Gabor滤波。提出了在二值图像上利用Gabor滤波器增强的方法,保留了Gabor滤波器连接纹线断线和滤除图像中噪声的优点,避免了Gabor滤波器对频率的敏感性问题,有效地提高了指纹特征的保持度。

During the preprocess of fingerprint images,Gabor filter shows prominent effect in image smoothing and segmentation. But Gabor filter is also very sensitive to the changes of ridge line direction and frequency,which leads to severe loses and changes of fingerprint feature in actual practice.To eliminate the defects,phase difference method are introduced in binarization, then Gabor filter is used for binary image enhancement.The proposed method not only reserves the Gabor filter＇s advantages,such as connection of broken line and noise reduction,but also avoids its limitation,such as more sensitivity to the changes of frequency.So,in binarization process,it can effectively preserve fingerprint features.