Efficient OCT Image Segmentation Using Neural Architecture Search

TL;DR

This paper introduces a neural architecture search framework that automates the design of a U-Net based model for retinal layer segmentation in OCT scans, achieving superior accuracy over traditional methods.

Contribution

The study presents a novel NAS approach incorporating U-Net for OCT segmentation, with a practical search strategy and pre-processing techniques to enhance image quality.

Findings

Achieved 95.4% mean IoU on OCT dataset

Attained 78.7% Dice coefficient

Outperformed human-designed architectures

Abstract

In this work, we propose a Neural Architecture Search (NAS) for retinal layer segmentation in Optical Coherence Tomography (OCT) scans. We incorporate the Unet architecture in the NAS framework as its backbone for the segmentation of the retinal layers in our collected and pre-processed OCT image dataset. At the pre-processing stage, we conduct super resolution and image processing techniques on the raw OCT scans to improve the quality of the raw images. For our search strategy, different primitive operations are suggested to find the down- & up-sampling cell blocks, and the binary gate method is applied to make the search strategy practical for the task in hand. We empirically evaluated our method on our in-house OCT dataset. The experimental results demonstrate that the self-adapting NAS-Unet architecture substantially outperformed the competitive human-designed architecture by achieving 95.4% in mean Intersection over Union metric and 78.7% in Dice similarity coefficient.