Abstract: To address the problem of segmenting the iris inaccurately in Hough transform when iris images are corrupted by the eyelids,eyelashes and deformation,an improved method of iris image segmentation based on image alignment is presented.The proposed method uses threshold and Hough transform to locate the center of pupil and applies Harris corner detection algorithm to estimate the left corner of eye.And it employs the robust alignment by sparse and low-rank decomposition algorithm to deal the labeled images with the two detected points,to make it have the low-rank feature.The proposed method uses edge detection and Hough transform method on the processed images to segment the iris accurately.Experimental results show that compared with Hough transform,this method can effectively remove the eyelids and eyelashes occlusion,and improve the accuracy of iris localization.

Key words: Hough transform, image alignment, iris image segmentation, Harris corner detection, low-rank, edge detection

摘要： 在传统的Hough变换虹膜图像分割方法中,虹膜图像易受到外界因素影响而造成分割不准确。针对该问题,提出一种虹膜图像分割方法。采用阈值法和Hough变化的方法检测出瞳孔中心,Harris角点检测法得出左眼角,使用稀疏和低秩分解的批量对齐算法对已分类标记的图像进行处理,使其具有低秩特性。对处理后的图像应用边缘检测和Hough变换相结合的方法实现人眼虹膜的分割。与传统的Hough变换方法进行实验对比,结果表明该方法能有效地去除因遮挡形成的噪声,从而提高虹膜定位的精确度。

关键词: Hough变换, 图像对齐, 虹膜图像分割, Harris角点检测, 低秩, 边缘检测

CLC Number: 

• TP391.41