Towards Ophthalmologist Level Accurate Deep Learning System for OCT\n Screening and Diagnosis

2018/12/12 by Mrinal Haloi, Haloi, Mrinal
Medicine · #68T45 #Artificial Intelligence (cs.AI) #COVID-19 diagnosis using AI #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Retinal Diseases and Treatments #Retinal Imaging and Analysis #Retinal and Optic Conditions

paper · pdf · doi:10.48550/arxiv.1812.07105

openalex publication_date 2018/12/12 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

In this work, we propose an advanced AI based grading system for OCT images.\nThe proposed system is a very deep fully convolutional attentive classification\nnetwork trained with end to end advanced transfer learning with online random\naugmentation. It uses quasi random augmentation that outputs confidence values\nfor diseases prevalence during inference. Its a fully automated retinal OCT\nanalysis AI system capable of pathological lesions understanding without any\noffline preprocessing/postprocessing step or manual feature extraction. We\npresent a state of the art performance on the publicly available Mendeley OCT\ndataset.\n

Related

• Deep learning achieves perfect anomaly detection on 108,308 retinal images including unlearned diseases
• Comparisonal study of Deep Learning approaches on Retinal OCT Image
• Retinal OCT disease classification with variational autoencoder regularization
• Enhancing Retinal Disease Classification from OCTA Images via Active Learning Techniques
• Less is more: Ensemble Learning for Retinal Disease Recognition Under Limited Resources