Mitigating Stethoscope-Induced Shortcuts in Respiratory Sound Classification under Federated Domain Generalization with Causality-Inspired Interventions

2026-05-28

AI Summary

This work addresses the limited generalization of existing respiratory sound classification models across diverse stethoscopes, which hinders deployment on unseen devices due to device-induced distribution shifts. To mitigate this, the study introduces causal intervention into a federated domain generalization framework and proposes a multimodal approach that disentangles device-specific style from pathological content through content-preserving style perturbation. The method further neutralizes metadata shortcuts by integrating counterfactual text augmentation, cross-client gradient alignment, and a language-audio pretrained model. Evaluated under a leave-one-device-out protocol on the ICBHI and SPRSound datasets, the proposed approach significantly outperforms current data augmentation and federated learning baselines, demonstrating enhanced robustness and generalization to previously unencountered stethoscopes.

Abstract

AI-driven respiratory sound classification (RSC) is promising for automated pulmonary disease detection, yet multi-site deployment is hindered by inter-stethoscope variability. We introduce a federated domain generalization (FedDG) formulation for RSC under stethoscope-induced device shifts, where clients use heterogeneous devices and the model is evaluated on unseen devices. Our empirical analysis shows that stethoscope-induced style and disease-specific content are tightly entangled, making deterministic style removal unreliable. In response, we propose a causality-inspired multimodal FedDG framework that combines: (i) a causality-inspired device style intervention network that performs content-preserving style perturbations, (ii) counterfactual text augmentation that neutralizes metadata shortcuts, and (iii) gradient alignment that facilitates device-invariant representations across clients. Built on a multimodal language-audio pretraining model, it outperforms conventional data augmentation and federated learning baselines in leave-one-device-out validation on ICBHI and SPRSound datasets. Code will be released upon publication.

Problem

• respiratory sound classification
• stethoscope-induced shortcuts
• federated domain generalization
• device shift
• causality

Innovation

• federated domain generalization
• causality-inspired intervention
• device-invariant representation
• counterfactual augmentation
• respiratory sound classification