一种基于支持向量机的自适应虹膜定位算法

A Self-Adaptive Iris Locating Algorithm Based on SVM

由于参数的限制和虹膜外边界两边低的对比度,使得Canny算子只能检测到位于某些频率段中的边缘,有时会漏掉一些比较模糊或者一些干扰较大的外边界.为了解决Canny算子进行虹膜定位过程中的鲁棒性问题,提出使用支持向量机(SVM)的自适应以改变Canny算子参数的方法.该算法通过傅立叶变换判断眼睛图像的质量,再使用SVM对眼睛图像分类,然后对于不同质量的眼睛图像采用不同参数的Canny算子检测边缘.实验结果表明该算法可以更加准确地检测到虹膜外边界,并有效地解决了虹膜定位中Canny算子的参数依赖问题.

Owing to the constraint of parameters and the low contrast between the two sides of the boundary, canny operator can only detect the border under some certain band of frequency and may omit some obscure edges or disturbed edges. To solve the robust problem of canny operator in iris locating, the paper presents a method to self-adaptively changes parameters of canny operator. The algorithm estimates the optical image quality using Fourier Transform, and uses Support Vector Machine （SVM） to classify the optical image, then detects edges using Canny Operator with different parameters to images with different quality. The result of experiment shows the method can detect the iris outer border precisely. The paper effectively solves parameters reliance of Canny Operator in iris locating.