See Through the Fog: Curriculum Learning with Progressive Occlusion in Medical Imaging

TL;DR

This paper introduces a curriculum learning approach with progressive occlusion to improve deep learning models' robustness in medical imaging, using novel occlusion synthesis methods and demonstrating significant accuracy gains.

Contribution

It presents a new curriculum learning framework with three innovative occlusion synthesis techniques to enhance model performance on occluded medical images.

Findings

Improved diagnostic accuracy on occluded images

Enhanced model robustness against occlusion

Effective training methodology demonstrated across datasets

Abstract

In recent years, deep learning models have revolutionized medical image interpretation, offering substantial improvements in diagnostic accuracy. However, these models often struggle with challenging images where critical features are partially or fully occluded, which is a common scenario in clinical practice. In this paper, we propose a novel curriculum learning-based approach to train deep learning models to handle occluded medical images effectively. Our method progressively introduces occlusion, starting from clear, unobstructed images and gradually moving to images with increasing occlusion levels. This ordered learning process, akin to human learning, allows the model to first grasp simple, discernable patterns and subsequently build upon this knowledge to understand more complicated, occluded scenarios. Furthermore, we present three novel occlusion synthesis methods, namely Wasserstein Curriculum Learning (WCL), Information Adaptive Learning (IAL), and Geodesic Curriculum Learning (GCL). Our extensive experiments on diverse medical image datasets demonstrate substantial improvements in model robustness and diagnostic accuracy over conventional training methodologies.