Faithful, Interpretable Chest X-ray Diagnosis with Anti-Aliased B-cos Networks

TL;DR

This paper enhances B-cos neural networks with anti-aliasing techniques to produce more faithful and interpretable chest X-ray diagnoses, maintaining accuracy while improving explanation clarity for clinical use.

Contribution

Introduces anti-aliasing strategies using FLCPooling and BlurPool to improve explanation quality of B-cos networks in medical imaging.

Findings

Improved explanation maps without aliasing artifacts.

Maintained diagnostic accuracy comparable to standard models.

Applicable to multi-class and multi-label chest X-ray diagnosis.

Abstract

Faithfulness and interpretability are essential for deploying deep neural networks (DNNs) in safety-critical domains such as medical imaging. B-cos networks offer a promising solution by replacing standard linear layers with a weight-input alignment mechanism, producing inherently interpretable, class-specific explanations without post-hoc methods. While maintaining diagnostic performance competitive with state-of-the-art DNNs, standard B-cos models suffer from severe aliasing artifacts in their explanation maps, making them unsuitable for clinical use where clarity is essential. In this work, we address these limitations by introducing anti-aliasing strategies using FLCPooling (FLC) and BlurPool (BP) to significantly improve explanation quality. Our experiments on chest X-ray datasets demonstrate that the modified $\text{B-cos}_\text{FLC}$ and $\text{B-cos}_\text{BP}$ preserve strong predictive performance while providing faithful and artifact-free explanations suitable for clinical application in multi-class and multi-label settings. Code available at: GitHub repository (url: https://github.com/mkleinma/B-cos-medical-paper).