公开(公告)号：US20170132793A1

公开(公告)日：2017-05-11

申请号：US15412838

申请日：2017-01-23

Applicant: Shenzhen Institutes of Advanced Technology

Inventor： Qingmao HU ,  Tianqiao ZHANG

IPC: G06T7/12 ,  G06T7/00 ,  G06T7/11 ,  G06T5/20

CPC classification number: G06T7/12 ,  G06T5/20 ,  G06T7/0012 ,  G06T7/11 ,  G06T7/149 ,  G06T2207/10101 ,  G06T2207/20016 ,  G06T2207/20024 ,  G06T2207/20104 ,  G06T2207/20116 ,  G06T2207/30041

Abstract: The present disclosure provides a method for acquiring a retina structure from an optical coherence tomographic image and a system thereof. The method comprises: calculating a region of interest (ROI) image of a source image; performing a Gaussian filtering of the ROI image; calculating a first estimated boundary position of a first layer boundary and a second estimated boundary position of a second layer boundary using a multi-resolution method; refining the first layer boundary and the second layer boundary respectively using a simplified active contour model according to the first estimated boundary position and the second estimated boundary position, to obtain a first initial boundary position of the first layer boundary and a second initial boundary position of the second layer boundary; smoothing the first initial boundary position and the second initial boundary position using a filtering method; acquiring segmentation positions of rest layer boundaries in the ROI image according to a segmented position of the first layer boundary and a segmented position of the second layer boundary. The present disclosure can improve the calculation speed and reduce the calculation complexity while maintaining high location accuracy.