Cost-Effective Active Learning for Melanoma Segmentation

TL;DR

This paper introduces a cost-effective active learning method for medical image segmentation that leverages dropout-based uncertainty estimation to reduce the need for extensive labeled datasets, improving training efficiency.

Contribution

It presents a practical active learning framework using dropout at test time for uncertainty modeling, tailored for medical image segmentation tasks.

Findings

Effective segmentation with limited labeled data

Uncertainty-based sample selection improves training

Open-source code available for reproducibility

Abstract

We propose a novel Active Learning framework capable to train effectively a convolutional neural network for semantic segmentation of medical imaging, with a limited amount of training labeled data. Our contribution is a practical Cost-Effective Active Learning approach using dropout at test time as Monte Carlo sampling to model the pixel-wise uncertainty and to analyze the image information to improve the training performance. The source code of this project is available at https://marc-gorriz.github.io/CEAL-Medical-Image-Segmentation/ .