Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentations

TL;DR

This paper explores the effectiveness of the Generalized Dice overlap as a loss function in deep learning for highly unbalanced medical image segmentation, addressing class imbalance issues.

Contribution

It introduces the use of the Generalized Dice overlap as a robust loss function for unbalanced segmentation tasks, improving performance over traditional methods.

Findings

Generalized Dice overlap improves segmentation accuracy in unbalanced data

It is less sensitive to learning rate tuning compared to other loss functions

Demonstrates effectiveness on 2D and 3D medical image segmentation tasks

Abstract

Deep-learning has proved in recent years to be a powerful tool for image analysis and is now widely used to segment both 2D and 3D medical images. Deep-learning segmentation frameworks rely not only on the choice of network architecture but also on the choice of loss function. When the segmentation process targets rare observations, a severe class imbalance is likely to occur between candidate labels, thus resulting in sub-optimal performance. In order to mitigate this issue, strategies such as the weighted cross-entropy function, the sensitivity function or the Dice loss function, have been proposed. In this work, we investigate the behavior of these loss functions and their sensitivity to learning rate tuning in the presence of different rates of label imbalance across 2D and 3D segmentation tasks. We also propose to use the class re-balancing properties of the Generalized Dice overlap, a known metric for segmentation assessment, as a robust and accurate deep-learning loss function for unbalanced tasks.