Abstract:<h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Background and Objectives</h3><p>Pathological myopia (PM) is the seventh leading cause of blindness, with a reported global prevalence up to 3%. Early and automated PM detection from fundus images could aid to prevent blindness in a world population that is characterized by a rising myopia prevalence. We aim to assess the use of convolutional neural networks (CNNs) for the detection of PM and semantic segmentation of myopia-induced lesions from fundus images on a recently introduced reference data set.</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Methods</h3><p>This investigation reports on the results of CNNs developed for the recently introduced Pathological Myopia (PALM) dataset, which consists of 1200 images. Our CNN bundles lesion segmentation and PM classification, as the two tasks are heavily intertwined. Domain knowledge is also inserted through the introduction of a new Optic Nerve Head (ONH)-based prediction enhancement for the segmentation of atrophy and fovea localization. Finally, we are the first to approach fovea localization using segmentation instead of detection or regression models. Evaluation metrics include area under the receiver operating characteristic curve (AUC) for PM detection, Euclidean distance for fovea localization, and Dice and F1 metrics for the semantic segmentation tasks (optic disc, retinal atrophy and retinal detachment).</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Results</h3><p>Models trained with 400 available training images achieved an AUC of 0.9867 for PM detection, 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.</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Conclusions</h3><p>We report a successful approach for a simultaneous classification of pathological myopia and segmentation of associated lesions. 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 glaucoma deep learning models, we envisage 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.</p>

Pathological Myopia detection from selective fundus image features

Automatic Detection Of Pathological Myopia And High Myopia On Fundus Images

Automatic detection of pathological myopia using machine learning

Multi-class classification of pathological myopia based on fundus photography

Development of deep learning-based detecting systems for pathologic myopia using retinal fundus images

A Curriculum Learning-Based Fully Automated System for Quantification of the Choroidal Structure in Highly Myopic Patients

Efficient Pyramid Channel Attention Network for Pathological Myopia Recognition

Automated detection of myopic maculopathy using five-category models based on vision outlooker for visual recognition

Deep Learning Approach for Automated Detection of Myopic Maculopathy and Pathologic Myopia in Fundus Images

Efficacy of a Deep Learning System for Screening Myopic Maculopathy Based on Color Fundus Photographs

MIU-Net: Advanced Multi-Scale Feature Extraction and Imbalance Mitigation for Optic Disc Segmentation

DBA-PMC: A Mutually Enhancing Dual-Branch Architecture for Pathologic Myopia and Myopic Maculopathy Classification

Self-supervised learning-enhanced deep learning method for identifying myopic maculopathy in high myopia patients

Open Fundus Photograph Dataset with Pathologic Myopia Recognition and Anatomical Structure Annotation

Automatic identification of myopic maculopathy related imaging features in optic disc region via machine learning methods

MyopiaDETR: End-to-end pathological myopia detection based on transformer using 2D fundus images

Smartphone App to Detect Pathological Myopia Using Spatial Attention and Squeeze‐Excitation Network as a Classifier and Segmentation Encoder

AI-Model for Identifying Pathologic Myopia Based on Deep Learning Algorithms of Myopic Maculopathy Classification and “Plus” Lesion Detection in Fundus Images

Automatic lesion segmentation and Pathological Myopia classification in fundus images

Use of multimodal dataset in AI for detecting glaucoma based on fundus photographs assessed with OCT: focus group study on high prevalence of myopia

Pathological myopia classification with simultaneous lesion segmentation using deep learning