裂隙灯显微图像角膜的快速、自动三维重建  

Fast-Auto Approach of Cornea Reconstruction based on slit-lamp Micrograph

作　　者：戴培山[1] 王博亮[2] 鞠颖[2] 程明[2] 汤井田[3] 

机构地区：[1]国防科学与技术大学电子科学与技术系,长沙410073 [2]厦门大学计算机系,厦门361005 [3]中南大学信息物理工程学院,长沙410083

出　　处：《系统仿真学报》2007年第5期1049-1052,共4页Journal of System Simulation

基　　金：国家自然科学基金项目(60371012);福建省科技重点项目(2005Y0018);卫生部联合基金项目(WKJ2005-2-001)

摘　　要：角膜病变大都与角膜的几何形状的改变密切相关,快速自动的显示角膜三维形状对角膜疾病的诊治具有重要意义。提出了一种基于裂隙灯显微图像的快速、自动的角膜分割、重建方法。基于蓝色分量水平投影累积量实现了角膜快速定位。采用颜色聚类分割出角膜部分。利用一种基于曲线拟合思想的方法去除大目标周围小噪声。对于病变严重的角膜我们采用面积最大的方法识别角膜,去除噪声。采用三次样条曲线插值角膜的内外表面。在基于模型的配准中,利用角膜的对称性,设计了配准模型。实现了面绘制的角膜重建。提出方法不需要人工交互,重建速度快。重建效果较好。Most corneal disease is closely related to cornea deform. A fast-auto approach of cornea reconstruction based on slit-lamp micrograph was designed Before segmentation of the image, a step of cornea orientation based on blue color histogram was added, greatly improving image process speed A color clustering method was used to segment the slit-lamp micrograph. A new method to remove noise besides big object was given. Cubic splines interpolation was used to get smooth corneal surface. Based on the symmetrical feature of corneal surface, a 3D model for the 2D-image sequence registration was designed. The cornea auto reconstruction was realized. The reconstruction speed is very fast and needs no interaction during the process.

关 键 词：裂隙灯显微图像 角膜 图像处理 三维重建 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]