A XAI-based Framework for Frequency Subband Characterization of Cough Spectrograms in Chronic Respiratory Disease
#XAI #Cough spectrogram #Chronic Obstructive Pulmonary Disease #Convolutional Neural Network #Occlusion maps #Frequency subband analysis #Time‑frequency representation #Medical AI tools
📌 Key Takeaways
- Development of an XAI‑based framework for cough spectrogram analysis.
- Utilization of a Convolutional Neural Network (CNN) on time‑frequency representations.
- Employment of occlusion maps to identify critical regions in spectrograms.
- Focus on chronic respiratory diseases, particularly Chronic Obstructive Pulmonary Disease (COPD).
- Aim to improve diagnostic accuracy and model interpretability.
📖 Full Retelling
Researchers have announced a new explainable artificial intelligence (XAI) framework designed to analyze cough sounds associated with chronic respiratory diseases, specifically Chronic Obstructive Pulmonary Disease (COPD). The study, published on arXiv (ID 2508.16237v2) in August 2025, reports the use of a Convolutional Neural Network (CNN) trained on time‑frequency representations of cough signals. By applying occlusion maps, the team can highlight diagnostically relevant regions within the spectrograms, offering insights into which frequency subbands most contribute to disease classification. The goal is to enhance the interpretability of machine‑learning models and support clinicians in making more informed diagnostic decisions for patients with COPD and potentially other chronic respiratory conditions.
🏷️ Themes
Explainable artificial intelligence, Healthcare diagnostics, Audio signal processing, Chronic respiratory disease, Neural network interpretability
Entity Intersection Graph
No entity connections available yet for this article.
Original Source
arXiv:2508.16237v2 Announce Type: replace-cross
Abstract: This paper presents an explainable artificial intelligence (XAI)-based framework for the spectral analysis of cough sounds associated with chronic respiratory diseases, with a particular focus on Chronic Obstructive Pulmonary Disease (COPD). A Convolutional Neural Network (CNN) is trained on time-frequency representations of cough signals, and occlusion maps are used to identify diagnostically relevant regions within the spectrograms. Th
Read full article at source