基于Markov随机场的改进的传统眼皮区域检测算法研究

目的]阐述改进的虹膜有效区域检测阶段中眼皮区域检测算法。方法]介绍传统的基于Radon变换的眼皮区域检测算法,针对其不能很好地消除眼皮阴影的缺点,提出基于Markov随机场理论眼部阴影检测算法。根据Markov随机场邻域的约束,使用MAP准则进行求解,利用K-mean求取先验信息,使用EM算法确定图像中每个像素的观测场模型,根据统计学理论的最优准则,得到能量函数的最小值,将眼皮阴影检测出来。结果]该研究提出的方法可以有效地去除虹膜有效区域检测阶段中眼皮阴影。结论]基于Markov随机场理论的眼皮阴影检测算法可以作为Radon眼皮检测的补充和改进。

The Improved Traditional Eyelid Region Detection Algorithm Based on Markov Random Field

Objective ] The aim was to elaborate the improved eyelid region detection algorithm during the ins eitectlve area aetectmn ptaase. Method] The traditional detection algorithm based on Radon transform eyelid region was introduced, the eye shadow detection algorithm based on Markov random field were summed up according to its disadvantage of not good for eliminating the eye shadow. According to the the neighborhood constraints of Markov random field, the MAP criteria was used to solve, the apriori information was obtained using K-mean algorithm, the observation field model of each pixel of image was confirmed on using EM algorithm, the minimum energy function was obtained according to the statistical theory of optimal criteria, and then the eyelid shadow was detected. Result] The detection algorithm proposed by this research eliminated the eyelid shadow during the iris effective area detection phase effectively. Conclusion] The eyelid shadow detection algorithm based on Markov random field theory can be used as the supplement and improvement of Radon eyelid detection.