Learning Non-Unique Segmentation with Reward-Penalty Dice Loss

TL;DR

This paper introduces the reward-penalty Dice loss (RPDL), a novel training objective for deep neural networks to effectively learn non-unique segmentation tasks, such as in medical applications with multiple valid annotations.

Contribution

The paper proposes RPDL, a new loss function that enhances learning of non-unique segmentation by emphasizing common regions and penalizing outside areas, addressing a gap in current segmentation methods.

Findings

RPDL improves segmentation accuracy by up to 18.4% on surgical datasets.

RPDL effectively learns from non-unique annotations in medical images.

The method outperforms traditional loss functions in non-unique segmentation tasks.

Abstract

Semantic segmentation is one of the key problems in the field of computer vision, as it enables computer image understanding. However, most research and applications of semantic segmentation focus on addressing unique segmentation problems, where there is only one gold standard segmentation result for every input image. This may not be true in some problems, e.g., medical applications. We may have non-unique segmentation annotations as different surgeons may perform successful surgeries for the same patient in slightly different ways. To comprehensively learn non-unique segmentation tasks, we propose the reward-penalty Dice loss (RPDL) function as the optimization objective for deep convolutional neural networks (DCNN). RPDL is capable of helping DCNN learn non-unique segmentation by enhancing common regions and penalizing outside ones. Experimental results show that RPDL improves the performance of DCNN models by up to 18.4% compared with other loss functions on our collected surgical dataset.