Stochastic Segmentation Networks: Modelling Spatially Correlated Aleatoric Uncertainty

TL;DR

This paper introduces stochastic segmentation networks (SSNs), a probabilistic approach that models spatially correlated aleatoric uncertainty in image segmentation, generating multiple plausible, coherent hypotheses efficiently, especially useful in ambiguous medical imaging scenarios.

Contribution

The paper presents SSNs, a novel method that models joint distributions over entire label maps using low-rank multivariate normal distributions, improving uncertainty estimation in segmentation tasks.

Findings

SSNs outperform state-of-the-art methods in modeling correlated uncertainty.

SSNs are more flexible and efficient than existing approaches.

SSNs generate multiple spatially coherent hypotheses effectively.

Abstract

In image segmentation, there is often more than one plausible solution for a given input. In medical imaging, for example, experts will often disagree about the exact location of object boundaries. Estimating this inherent uncertainty and predicting multiple plausible hypotheses is of great interest in many applications, yet this ability is lacking in most current deep learning methods. In this paper, we introduce stochastic segmentation networks (SSNs), an efficient probabilistic method for modelling aleatoric uncertainty with any image segmentation network architecture. In contrast to approaches that produce pixel-wise estimates, SSNs model joint distributions over entire label maps and thus can generate multiple spatially coherent hypotheses for a single image. By using a low-rank multivariate normal distribution over the logit space to model the probability of the label map given the image, we obtain a spatially consistent probability distribution that can be efficiently computed by a neural network without any changes to the underlying architecture. We tested our method on the segmentation of real-world medical data, including lung nodules in 2D CT and brain tumours in 3D multimodal MRI scans. SSNs outperform state-of-the-art for modelling correlated uncertainty in ambiguous images while being much simpler, more flexible, and more efficient.