Training Multi-organ Segmentation Networks with Sample Selection by Relaxed Upper Confident Bound

TL;DR

This paper introduces a novel sample selection method called Relaxed Upper Confident Bound (RUCB) for training multi-organ segmentation networks, improving performance by balancing hard and less frequent samples while mitigating annotation errors.

Contribution

The paper proposes the RUCB sample selection policy, which enhances training of segmentation networks by exploiting a broader range of hard samples and reducing annotation error impact.

Findings

RUCB improves segmentation accuracy on abdominal CT scans.

The method effectively balances hard sample exploitation and exploration.

Performance gains are significant compared to baseline sampling strategies.

Abstract

Deep convolutional neural networks (CNNs), especially fully convolutional networks, have been widely applied to automatic medical image segmentation problems, e.g., multi-organ segmentation. Existing CNN-based segmentation methods mainly focus on looking for increasingly powerful network architectures, but pay less attention to data sampling strategies for training networks more effectively. In this paper, we present a simple but effective sample selection method for training multi-organ segmentation networks. Sample selection exhibits an exploitation-exploration strategy, i.e., exploiting hard samples and exploring less frequently visited samples. Based on the fact that very hard samples might have annotation errors, we propose a new sample selection policy, named Relaxed Upper Confident Bound (RUCB). Compared with other sample selection policies, e.g., Upper Confident Bound (UCB), it exploits a range of hard samples rather than being stuck with a small set of very hard ones, which mitigates the influence of annotation errors during training. We apply this new sample selection policy to training a multi-organ segmentation network on a dataset containing 120 abdominal CT scans and show that it boosts segmentation performance significantly.