Inherent Brain Segmentation Quality Control from Fully ConvNet Monte Carlo Sampling

TL;DR

This paper presents a Bayesian neural network approach that uses model uncertainty to automatically assess and improve the quality control of brain segmentation, enhancing clinical and large-scale data analysis.

Contribution

It introduces novel structure-wise uncertainty metrics derived from Monte Carlo dropout samples for effective segmentation quality assessment.

Findings

Uncertainty metrics strongly correlate with segmentation accuracy.

Group analysis with uncertainty yields effect sizes closer to manual annotations.

The approach enables automated quality control in large data repositories.

Abstract

We introduce inherent measures for effective quality control of brain segmentation based on a Bayesian fully convolutional neural network, using model uncertainty. Monte Carlo samples from the posterior distribution are efficiently generated using dropout at test time. Based on these samples, we introduce next to a voxel-wise uncertainty map also three metrics for structure-wise uncertainty. We then incorporate these structure-wise uncertainty in group analyses as a measure of confidence in the observation. Our results show that the metrics are highly correlated to segmentation accuracy and therefore present an inherent measure of segmentation quality. Furthermore, group analysis with uncertainty results in effect sizes closer to that of manual annotations. The introduced uncertainty metrics can not only be very useful in translation to clinical practice but also provide automated quality control and group analyses in processing large data repositories.