随机比较Gabor滤波的可撤销掌纹识别算法

Cancelable Palmprint Verification Using Gabor Based Random Measure

针对掌纹生物特征识别的模板安全保护,本文研究并提出一种Gabor滤波与混沌矩阵进行随机比较的可撤销掌纹识别方法。首先使用Gabor滤波器与经过预处理的掌纹图像卷积;然后将滤波得到的模板与混沌矩阵进行随机比较,得到可撤销的二值模板;最后根据计算模板之间的汉明距离归类匹配。通过掌纹数据库实验表明,该算法在保护掌纹图像模板安全性的同时,提高了识别效果。

This paper proposed a novel cancelable palmprint verification to deal with the security and privacy issues of template protection. The palmprint template is generated by measuring the Gabor based palmprint image with a chaotic random matrix. Firstly, The Gabor is employed to convolute with preprocessed palmprint image. Then, the Gabor based image is measured via random matrix and the obtained binary template is the cancelable template. The experimental results of verification on Hong Kong PolyU Palmprint Database show that the proposed approach not only protects the palmprint template but is also able to increase the template discriminability.