虹膜图像的分段检测眼睑算法提取及其图像清晰度评价研究  

Iris images＇ extraction based on the algorithm for segment detection of eyelid and their clarity assessment

作　　者：宋斐[1] 王金星[2] 马宜青[1] 张倩[1] 保文星[3] 

机构地区：[1]宁夏医科大学理学院,银川750004 [2]宁夏医科大学附属医院,银川750004 [3]北方民族大学计算机科学与工程学院,银川750001

出　　处：《高技术通讯》2011年第4期351-355,共5页Chinese High Technology Letters

基　　金：国家科技支撑计划（2007BAD33803）资助项目.

摘　　要：针对虹膜图像的提取和清晰度分析问题，研究了基于分段检测眼睑的虹膜图像采集及对所采图像的清晰度的评价，进而提出了一种新的解决方法。所提方法使用分段检测的眼睑分割算法提取虹膜边缘，使用倒谱域函数评价图像的清晰度。分段检测方法可经过圆弧参数的分段拟合、动态聚类，快速准确地分割出虹膜图像中的上下眼睑，倒谱域函数评价法可大大减少运算时间，提高抗噪性和可靠性。实验结果证明，该方法能有效地提高虹膜图像的识别率，且对整体识别时间影响很小，因而更适合用作眼睛图像的质量评价。On the issue of extraction of iris images and analysis of their clarity, this paper studies extracting iris images based on segment detection of eyelid and assessing the clarity of the images, and puts forward a new solution in this regard. In the solution, the edge of the iris can be extracted by using the algorithm for segment detection of eyelid, and the image clarity can be evaluated by using the cepstrum function. Through segment fitting and dynamic clustering of the circular parameter, the upper and lower eyelids in the iris images can be segmented in a fast and accurate manner. The cepstrum function method can significantly save time and improve noise immunity and reliability. The results of experiments have shown that the new method can improve the accuracy of iris recognition effectively, and it has little effect on the total recognition time; so it is more suitable for quality assessment of eyes image.

关 键 词：分段检测算法 眼睑分割 图像处理 倒谱域 清晰度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]