Disentangling Human Error from the Ground Truth in Segmentation of Medical Images

TL;DR

This paper introduces a novel CNN-based method to disentangle human errors from true labels in medical image segmentation, improving accuracy by modeling annotator reliability from noisy, diverse annotations.

Contribution

The work presents a joint learning approach that estimates annotator reliability and true labels simultaneously using coupled CNNs, addressing label noise in medical image segmentation.

Findings

Outperforms existing methods on multiple medical datasets.

Effectively models complex spatial annotator errors.

Excels especially with limited annotations and high disagreement.

Abstract

Recent years have seen increasing use of supervised learning methods for segmentation tasks. However, the predictive performance of these algorithms depends on the quality of labels. This problem is particularly pertinent in the medical image domain, where both the annotation cost and inter-observer variability are high. In a typical label acquisition process, different human experts provide their estimates of the "true" segmentation labels under the influence of their own biases and competence levels. Treating these noisy labels blindly as the ground truth limits the performance that automatic segmentation algorithms can achieve. In this work, we present a method for jointly learning, from purely noisy observations alone, the reliability of individual annotators and the true segmentation label distributions, using two coupled CNNs. The separation of the two is achieved by encouraging the estimated annotators to be maximally unreliable while achieving high fidelity with the noisy training data. We first define a toy segmentation dataset based on MNIST and study the properties of the proposed algorithm. We then demonstrate the utility of the method on three public medical imaging segmentation datasets with simulated (when necessary) and real diverse annotations: 1) MSLSC (multiple-sclerosis lesions); 2) BraTS (brain tumours); 3) LIDC-IDRI (lung abnormalities). In all cases, our method outperforms competing methods and relevant baselines particularly in cases where the number of annotations is small and the amount of disagreement is large. The experiments also show strong ability to capture the complex spatial characteristics of annotators' mistakes.