基于超高分辨率光学相干断层扫描( OCT )图像提出了视网膜层次自动分割的方法。利用图论和基于动态规划的最短路算法，自动寻找出准确的视网膜八层结构，验证了算法的准确性和可靠性。自动和手动探测视网膜各层边界的结果具有较强的可比性，同时此算法适用于青光眼、高度近视眼、视网膜色素变性等患者黄斑区视网膜层状结构的分割。视网膜层状结构自动分割方法为临床诊断和治疗提供了定量分析的工具。%To evaluate the automated segmentation algorithm for detection of intra - retinal layers to process images obtained from ultra- high resolution optical coherence tomography ( OCT ) . Graph theory and the shortest path search based on dynamic programming were applied to automatically segment the 8 intra - retinal layers. We experimentally verified the accuracy and reliability of the algorithm. The results showed that the intra-retinal layer boundaries between automated and manual segmentations matched well. The algorithm successfully segmented the intra- retinal layers in glaucoma, high myopia, and retinitis pigmentosa patients. The proposed automatic segmentation for intra-retinal layers provides a promising tool for quantitative analysis in clinical diagnosis and treatment.
展开▼
