基于Radon变换的多尺度虹膜特征提取算法

A multi-scale iris feature extraction algorithm based on radon transform

系统分析了基于虹膜特征的身份识别系统。针对典型预处理特点,设计了基于Radon变换的新特征提取算法。该算法在保持旋转、平移和尺度缩放不变性的同时,利用标准化方差加权的绝对距离,使识别准确度达到了93%。这种算法对于其它不规则图像的分析也具有启发意义。

An identification system based on iris feature is analyzed in this paper. Aiming to the typical characteristic of image preprocessing algorithms,a new figure extraction algorithm based on Radon transform is presented, This method makes the corrent recognition rate reach 93 % by using the standardized variance weighted absolute distance,meanwhile keeping the in variance of rotatin,translation and scale, The algorithm is also of heuristic significance for irregular images analysis.