D Model Reconstruction of Blood Vessels in The Retina with Tubular Structure
D Model Reconstruction of Blood Vessels in The Retina with Tubular Structure
复制标题
管状结构视网膜血管D模型重建
DOI：
10.15676/ijeei.2015.7.4.14
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
H. Belmabrouk
中科院分区：
文献类型：
--
作者：
G. Hichem;Fraj Chouchene;H. Belmabrouk
我们在本文中提出了人类视网膜血管 3D 模型重建的几何解释。所提出的方法将船舶跟踪算法的方法与广义圆柱体的连续模型的使用相结合。该模型由两个几何对象组成：生成器轴 H 和表面 S，表示为附加到该轴的一堆轮廓图。我们提出的监测血管分割算法是通过使用三种插值方法编写的，具有平滑的生成器轴 H。首先，从人类视网膜图像中提取血管的图形拓扑。这是在图像骨架化之后完成的，然后是像素分类（终止点、分叉点和内部点）以及分支曲线的确定。其次，我们使用三种方法来近似血管以获得平滑的曲线（线性、三次样条和最小二乘线性插值）。然后，使用右广义圆柱状态模型 (RGC-sm) 从血管图像构建 3D 模型。最后，在最后一部分中，我们展示了血管组的分割和 3D 重建结果。
We present in this paper a geometric interpretation of the 3D model reconstruction of the blood vessel of the human retina. The proposed method combines the approach of the vessel tracking algorithm with the use of a continuous model of generalized cylinder. The model is composed of two geometric objects: a generator axis H and a surface S, represented as a stack of contours plans attached to this axis. The segmentation algorithms by monitoring vessels that we propose is written by the use of three interpolation method for have a smooth generator axis H. First, the graph topology of the blood vessel is extracted from the human retina image. This is done after the skeletonisation of the image, then, the pixel classification (termination points and bifurcation points and inner points) and determination of the branches curve. Secondly, we use three methods of the approximation of the vessels for have a smooth curve (linear, cubic splines and least square linear interpolation). Afterwards, the construction of a 3D model from a vascular image is done by using the Right Generalized Cylinder State Model (RGC-sm). Finally, in a last part, we present results of segmentation and 3D reconstruction of a blood vessel group.