裂隙灯显微图像角膜的快速、自动三维重建

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。

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

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.