ClaRet -- A CNN Architecture for Optical Coherence Tomography

TL;DR

This paper introduces ClaRet, a CNN architecture leveraging transfer learning with VGG-19, to improve classification of retinal tears in OCT scans, demonstrating significant performance gains over baseline models.

Contribution

The paper presents a novel block-based CNN architecture, ClaRet, tailored for OCT scan classification, enhancing feature extraction and accuracy in detecting retinal tears.

Findings

Achieved higher accuracy than baseline models

Effective use of transfer learning with VGG-19

Improved feature extraction with custom blocks

Abstract

Optical Coherence Tomography is a technique used to scan the Retina of the eye and check for tears. In this paper, we develop a Convolutional Neural Network Architecture for OCT scan classification. The model is trained to detect Retinal tears from an OCT scan and classify the type of tear. We designed a block-based approach to accompany a pre-trained VGG-19 using Transfer Learning by writing customised layers in blocks for better feature extraction. The approach achieved substantially better results than the baseline we initially started out with.