一种指纹奇异点区域图像增强算法

指纹图像奇异点附近区域的增强一直是指纹图像增强的难点,针对Separable Gabor滤波会破坏指纹邻近奇异点区域的纹线结构,方向傅里叶滤波在一般区域修复指纹纹线效果不明显这一问题,本文融合两种算法的优势,提出一种新的滤波方法(FS-Gabor)。先对指纹图像进行预处理,得到指纹的方向、频率信息和掩膜信息。接着找出指纹图像的奇异点,并在奇异点附近标记出一定大小区域。最后根据像素点的位置采用不同的滤波方法。同时,本文提出了一种改进的指纹图像频率估计方法,扩大了指纹图像有效区域面积。实验结果表明,经本文方法滤波的指纹图像的EER(Equal Error Rate)比方向傅里叶滤波低26%,比Separable Gabor低49%。

Fingerprint singularity region image enhancement algorithm

Enhancement of fingerprint image near the singularity point is a big challenge in the fingerprint image enhancement. However, the ridge structure near the singular point of the fingerprint can be destroyed by the current Separable Gabor filter algorithm. And the performance in repairing the fingerprint ridge is bad by the current directional Fourier filter algorithm. These result in the bad performance in fingerprint image enhancement of the current Separable Gabor filter algorithm and Fourier filter algorithm. In order to improve the performance in the fingerprint image enhancement, this paper combines the advantages of the two algorithms and proposes the Fourier Separable Gabor algorithm (FS-Gabor). In the proposed FS-Gabor algorithm, the fingerprint image is preprocessed to obtain the fingerprint direction, frequency information and mask information firstly. Then in order to repaire the fingerprint ridge, the algorithm marks the area near the singular point as the aim filter area in the fingerprint image. Finally, different filtering methods are chosen to obtain the aim fingerprint image according to the position of the pixel. In addition, in order to expand the effective area of fingerprint image, an improved fingerprint image frequency estimation method is proposed in the FS-Gabor algorithm. According to the experiment results, it is indicated that the EER of fingerprint image filtered by the FS-Gabor algorithm is 26% lower than the directional Fourier filter algorithm and 49% lower than the Separable Gabor algorithm.