Estimation System using 3D Boundary Following on Retinal 3D OCT Image

本研究是利用高速光学相干层析成像单元开发一种新型的智能视网膜光学相干层析成像图像处理系统的研究。(1)构建了一种基于3D边界跟踪法和区域生长法的视网膜三维OCT图像分析系统，该系统能够自动估计视网膜组织上的异常面积和体积。(2)该系统是一个具有较高可用性的OCT图像软件系统，并与医生使用临床3D OCT图像进行对比，验证了对异常区域的提取准确率(老年性黄斑变性89%、黄斑变性73%、黄斑变性73%、总准确率79%)。

This research is a research on the development of a new intelligent image processing system on retinal Optical Coherence Tomography (OCT) image using high speed OCT unit. We obtained the following research results.(1) We constructed a new image analysis system including 3D boundary following method and region growing method on retinal 3D OCT image, and it can estimate the abnormal area and volume on a retinal tissue automatically.(2) This system is a software system with a high usability interface on OCT image, and we confirmed the extraction accuracy (89% for age related macular degeneration, 73% for drusen, 79% for total) of an abnormal area in comparison with medical doctors using the clinical 3D OCT images.