一种快速的视网膜光学相干图像分割方法

在医学领域，黄斑厚度可以用来量化糖尿病黄斑水肿和年龄相关性黄斑变性等疾病,临床上通常使用光学相干断层扫描的影像技术来获取黄斑图像。但现有的黄斑图像分割方法运算速度较慢，阻碍了其临床使用。本文提出一种新的基于多分辨率及水平集的黄斑图像分割方法，首先使用高斯滤波对原始图像按行进行滤波，再运用多分辨率方法获取图像初始局部轮廓，最后使用水平集方法可以快速获取黄斑图像的中间轮廓，得到最终的图像分割结果。通过在311幅黄斑图像的仿真实验对比，本文方法在边缘检测结果和运算速度上比传统方法有很大改进。

Fast Retinal Segmentation of Optical Coherence Tomography Images

Under the framework of computer-aided diagnosis, optical coherence tomography (OCT) has become an established ocular imaging technique to quantify diabetic macular edema and age-related macular degeneration (AMD) by measuring the macular thickness. However, existing methods have a high computational cost which hinders their clinical use. In this paper, a fast and robust method to measure macular thickness using spectral-domain optical coherence tomography images is presented. It consists of three steps. First, owing to the retina exhibiting a layered structure, two-dimensional image segmentation is converted to one-dimensional positioning to reduce computational cost. Then, multi-resolutions are employed to derive the initial contours, which can speed up furthermore. Finally, a fast level set approach is explored for obtaining the final segmentation. The performance of the approach has been validated against 311 macular images. Compared with other segmentation methods on OCT retina images, the proposed approach is both robust and fast.