Choroid Segmentation of Retinal OCT Images Based on CNN Classifier and <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" id="M1"> <a:msub> <a:mrow> <a:mi>l</a:mi> </a:mrow> <a:mrow> <a:mn>2</a:mn> </a:mrow> </a:msub> </a:math>-<c:math xmlns:c="http://www.w3.org/1998/Math/MathML" id="M2"> <c:msub> <c:mrow> <c:mi>l</c:mi> </c:mrow> <c:mrow> <c:mi>q</c:mi> </c:mrow> </c:msub> </c:math> Fitter
Fang He, Rachel Ka Man Chun, Zicheng Qiu, Shijie Yu, Yun Shi, Chi Ho To, Xiaojun Chen
Abstract
Optical coherence tomography (OCT) is a noninvasive cross-sectional imaging technology used to examine the retinal structure and pathology of the eye. Evaluating the thickness of the choroid using OCT images is of great interests for clinicians and researchers to monitor the choroidal thickness in many ocular diseases for diagnosis and management. However, manual segmentation and thickness profiling of choroid are time-consuming which lead to low efficiency in analyzing a large quantity of OCT images for swift treatment of patients. In this paper, an automatic segmentation approach based on convolutional neural network (CNN) classifier and <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" id="M3"> <a:msub> <a:mrow> <a:mi>l</a:mi> </a:mrow> <a:mrow> <a:mn>2</a:mn> </a:mrow> </a:msub> </a:math> - <c:math xmlns:c="http://www.w3.org/1998/Math/MathML" id="M4"> <c:msub> <c:mrow> <c:mi>l</c:mi> </c:mrow> <c:mrow> <c:mi>q</c:mi> </c:mrow> </c:msub> </c:math> ( <e:math xmlns:e="http://www.w3.org/1998/Math/MathML" id="M5"> <e:mn>0</e:mn> <e:mo><</e:mo> <e:mi>q</e:mi> <e:mo><</e:mo> <e:mn>1</e:mn> </e:math> ) fitter is presented to identify boundaries of the choroid and to generate thickness profile of the choroid from retinal OCT images. The method of detecting inner choroidal surface is motivated by its biological characteristics after light reflection, while the outer chorioscleral interface segmentation is transferred into a classification and fitting problem. The proposed method is tested in a data set of clinically obtained retinal OCT images with ground-truth marked by clinicians. Our numerical results demonstrate the effectiveness of the proposed approach to achieve stable and clinically accurate autosegmentation of the choroid.