Adaptive Differential Denoising for Respiratory Sounds Classification

2025/06/03 by Gaoyang Dong, Zhicheng Zhang, Dong, Gaoyang +5 · 1 citation
Computer Science · Health Professions · Medicine · #Audio and Speech Processing (eess.AS) #FOS: Computer and information sciences #FOS: Electrical engineering #Music and Audio Processing #Noise Effects and Management #Phonocardiography and Auscultation Techniques #Sound (cs.SD) #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.2506.02505

openalex publication_date 2025/06/03 · openalex created_date 2025/10/14 · openalex updated_date 2026/07/28

Abstract

Automated respiratory sound classification faces practical challenges from background noise and insufficient denoising in existing systems. We propose Adaptive Differential Denoising network, that integrates noise suppression and pathological feature preservation via three innovations: 1) Adaptive Frequency Filter with learnable spectral masks and soft shrink to eliminate noise while retaining diagnostic high-frequency components; 2) A Differential Denoise Layer using differential attention to reduce noise-induced variations through augmented sample comparisons; 3) A bias denoising loss jointly optimizing classification and robustness without clean labels. Experiments on the ICBHI2017 dataset show that our method achieves 65.53% of the Score, which is improved by 1.99% over the previous sota method. The code is available in https://github.com/deegy666/ADD-RSC

Citations

• BTS: Bridging Text and Sound Modalities for Metadata-Aided Respiratory Sound Classification
• RepAugment: Input-Agnostic Representation-Level Augmentation for Respiratory Sound Classification
• Masked Modeling Duo: Towards a Universal Audio Pre-training Framework
• Stethoscope-guided Supervised Contrastive Learning for Cross-domain Adaptation on Respiratory Sound Classification
• Multi-View Spectrogram Transformer for Respiratory Sound Classification
• Adversarial Fine-tuning using Generated Respiratory Sound to Address Class Imbalance
• Adaptive Frequency Filters As Efficient Global Token Mixers
• MossFormer: Pushing the Performance Limit of Monaural Speech Separation using Gated Single-Head Transformer with Convolution-Augmented Joint Self-Attentions
• Example-based Explanations with Adversarial Attacks for Respiratory Sound Analysis
• Lung Sound Classification Using Co-tuning and Stochastic Normalization
• RespireNet: A Deep Neural Network for Accurately Detecting Abnormal Lung Sounds in Limited Data Setting
• Pretraining Respiratory Sound Representations using Metadata and Contrastive Learning

Cited by

• Geometry-Aware Optimization for Respiratory Sound Classification: Enhancing Sensitivity with SAM-Optimized Audio Spectrogram Transformers

Related

• Disentangling Dual-Encoder Masked Autoencoder for Respiratory Sound Classification
• A Self-Attention-Driven Deep Denoiser Model for Real Time Lung Sound Denoising in Noisy Environments
• Robust Deep Learning Frameworks for Acoustic Scene and Respiratory Sound Classification
• Adversarial Fine-tuning using Generated Respiratory Sound to Address Class Imbalance
• Geometry-Aware Optimization for Respiratory Sound Classification: Enhancing Sensitivity with SAM-Optimized Audio Spectrogram Transformers