基于滤波和融合的指纹图像增强算法

指纹图像增强对于提高细节特征提取的准确率乃至整个自动指纹识别系统的性能都
具有重要的意义。大部分的指纹图像增强算法都需计算指纹图像的纹线局部方向，本文给出了一种指纹图像增强的新方法，该方法不计算指纹图像局部区域内的方向，而将指纹图像按8 个不同方向滤波，得到8 幅滤波后的图像，再利用小波变换对8 幅图像进行融合得到增强后的图像。实验表明，该算法有效的提高了指纹图像的质量。
关键词：指纹识别；图像增强；小波分析；图像融合
Abstract: Fingerprint image enhancement is important to feature extraction and the performance of fingerprint identification/verification systems rely. Most of the algorithms for fingerprint image enhancement depend on local orientation of fingerprint image. In this paper, an effective algorithm for fingerprint image enhancement which need not calculate the local orientation of fingerprint image is developed. First, the fingerprint image is filtered with eight different orientations. Then the eight filtered images are fused to get the enhanced image. The results of experiment indicate the algorithm improves the fingerprint images quality effectively.
Key words: Fingerprint verification; Image enhancement; Wavelet transform; Image fusion