摘要
针对虹膜图像中的眼睑、眼睫毛和光斑等噪声影响虹膜识别性能,提出一种虹膜分割中的噪声检测方法,在数学形态学的基础上,该方法采用边缘检测与最小二乘法相结合的方法拟合确定眼睑遮挡部分,利用阈值法来检测上眼睫毛和光斑。实验结果表明,该方法可准确而有效地在虹膜分割中检测和剔除这些噪声,是一种有效和可行的虹膜图像噪声检测算法。
In view of iris noise of the eyelid, eyelash and reflection spot affect performance of iris recognition, a method of noise detection about iris segmentation is proposed. Based on Mathematical morphology, the proposed method uses the least squares method combined with edge detection to determine eyelid occlusion regions, then, eyelash and reflection spot is detected with threshold method. The experiment indicated that this method may detect and suppress these noises accurately and effectively, it is an effective and feasible noise detection algorithm of iris image.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第10期2447-2449,共3页
Computer Engineering and Design
基金
国家自然科学基金项目(60472006)
广东韶关学院校级科研基金项目(2007年220号文件)
关键词
虹膜分割
检测
眼睑
睫毛
最小二乘法
阈值法
iris segmentation
detection
eyelid
eyelash
least squaresmethod
threshold method
