Pathological myopia classification with simultaneous lesion segmentation using deep learning

TL;DR

This paper develops convolutional neural networks for classifying pathological myopia and segmenting lesions in fundus images, achieving high accuracy and localization precision, and introduces an ONH-based prediction enhancement.

Contribution

It presents a novel deep learning approach with ONH-based prediction enhancement for simultaneous lesion segmentation and myopia classification on the PALM dataset.

Findings

Achieved AUC of 0.9867 for myopia classification

Dice scores over 0.93 for optic disc segmentation

Fovea localization Euclidean distance of 58.27 pixels

Abstract

This investigation reports on the results of convolutional neural networks developed for the recently introduced PathologicAL Myopia (PALM) dataset, which consists of 1200 fundus images. We propose a new Optic Nerve Head (ONH)-based prediction enhancement for the segmentation of atrophy and fovea. Models trained with 400 available training images achieved an AUC of 0.9867 for pathological myopia classification, and a Euclidean distance of 58.27 pixels on the fovea localization task, evaluated on a test set of 400 images. Dice and F1 metrics for semantic segmentation of lesions scored 0.9303 and 0.9869 on optic disc, 0.8001 and 0.9135 on retinal atrophy, and 0.8073 and 0.7059 on retinal detachment, respectively. Our work was acknowledged with an award in the context of the "PathologicAL Myopia detection from retinal images" challenge held during the IEEE International Symposium on Biomedical Imaging (April 2019). Considering that (pathological) myopia cases are often identified as false positives and negatives in classification systems for glaucoma, we envision that the current work could aid in future research to discriminate between glaucomatous and highly-myopic eyes, complemented by the localization and segmentation of landmarks such as fovea, optic disc and atrophy.