基于多通道Gabor滤波和特征融合的虹膜识别方法

Iris recognition based on multichannel Gabor filtering and feature fusion

多通道Gabor滤波器提取的虹膜特征具有冗余信息并存在部分非有效特征,针对此问题提出了改进方法。对同尺度不同方向的Gabor特征,利用幅值信息进行融合,对融合后特征进行相位编码,并运用海明距离匹配。这样,既保证了高识别性能,又将虹膜特征码压缩为传统方法的1/2,可提高匹配速度,并节约存储空间。还提出一种虹膜图像质量评价方法,可有效鉴别不适于识别的低质量虹膜图像。在CASIA和UBIRIS虹膜库的实验结果表明该方法是有效的。

To reduce redundancy of multichannel Gabor features, an approach based on feature-level fusion is proposed. Multiple Gabor features in the same scale with different orientations are fused by using the magnitude information, and the iris codes are generated based on the phase information. The similarity of two iris codes is measured by their hamming distance. Compared with traditional non-fusion approach, the proposed approach has the same high recognition performance, but the size of the iris codes in the proposed approach is only a half of the traditional one. In addition, a method for iris image quality estimate is presented, which can discriminate the images unsuitable for iris recognition. The experimental results show that the proposed approach is effective.