SUPER-Net: Trustworthy Image Segmentation via Uncertainty Propagation in Encoder-Decoder Networks

TL;DR

SUPER-Net introduces a Bayesian image segmentation framework that propagates uncertainty through encoder-decoder networks, enhancing robustness and providing pixel-wise confidence estimates without costly sampling.

Contribution

It presents a novel Taylor series-based uncertainty propagation method in Bayesian segmentation, improving reliability and efficiency over existing approaches.

Findings

Outperforms state-of-the-art models in noisy and adversarial conditions

Generates accurate pixel-wise uncertainty maps for reliability assessment

Eliminates the need for Monte Carlo sampling in uncertainty estimation

Abstract

Deep Learning (DL) holds great promise in reshaping the industry owing to its precision, efficiency, and objectivity. However, the brittleness of DL models to noisy and out-of-distribution inputs is ailing their deployment in sensitive fields. Current models often lack uncertainty quantification, providing only point estimates. We propose SUPER-Net, a Bayesian framework for trustworthy image segmentation via uncertainty propagation. Using Taylor series approximations, SUPER-Net propagates the mean and covariance of the model's posterior distribution across nonlinear layers. It generates two outputs simultaneously: the segmented image and a pixel-wise uncertainty map, eliminating the need for expensive Monte Carlo sampling. SUPER-Net's performance is extensively evaluated on MRI and CT scans under various noisy and adversarial conditions. Results show that SUPER-Net outperforms state-of-the-art models in robustness and accuracy. The uncertainty map identifies low-confidence areas affected by noise or attacks, allowing the model to self-assess segmentation reliability, particularly when errors arise from noise or adversarial examples.