EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Respiratory Disease Classification Using NMF-Enhanced Log-Mel Spectrograms and Convolutional Recurrent Neural Networks

Bowen Han, Wei Quan iD, Bogdan Matuszewski, Dennis Corbett

Sensors · 04.07.2026 · 26 (13) · 4268

DOI: 10.3390/s26134268 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 04.07.2026. Verfasst von Bowen Han und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Engineering and Computing, University of Lancashire, Preston PR1 2HE, UK. Hauptthemen laut MEDLINE: Respiratory Sounds, Respiratory Tract Diseases. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: University of Lancashire.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Respiratory disease classification using lung sound recordings remains challenging due to signal interference, heterogeneous acquisition conditions, and substantial overlap among clinically related acoustic patterns. This study presents a framework for respiratory disease classification using NMF-enhanced log-mel spectrograms and deep neural classifiers. Respiratory sound recordings from two publicly available datasets were harmonized into a unified label space comprising Asthma, Bronchiectasis, Bronchiolitis, COPD, Healthy, Pneumonia and URTI. Following signal standardization and fixed-length segmentation, a non-negative matrix factorization (NMF)-based enhancement stage was applied to increase the salience of respiratory components prior to log-mel spectrogram generation. The proposed classifier was a convolutional recurrent neural network (CRNN) that combined convolutional feature extraction, bidirectional recurrent modelling, and attention-based temporal aggregation. For comparison, RDLINet, a conventional CNN, ResNet, and a YOLO-style backbone were implemented under the same preprocessing and training framework. Experimental results demonstrated that the proposed CRNN achieved the best overall performance, attaining 96.14 ± 0.50% accuracy and 94.05 ± 1.21% Macro-F1 on the unified seven-class cohort. Class-wise analysis, confusion-matrix evaluation, and output-space visualization further showed that the CRNN provided more balanced recognition across disease categories and clearer class separation than competing architectures. These findings indicate that NMF-enhanced spectro-temporal modelling combined with convolutional recurrent learning offers an effective approach for automated multi-class respiratory disease classification.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Respiratory SoundsRespiratory Tract DiseasesHumansConvolutional Neural NetworksRecurrent Neural NetworksSignal Processing, Computer-AssistedNeural Networks, ComputerAlgorithmsSound Spectrography

Stichwörter der Autor:innen

computer-assisted auscultationconvolutional recurrent neural networklog-mel spectrogramnon-negative matrix factorizationrespiratory disease classification

Autor:innen und Institutionen

  • School of Engineering and Computing, University of Lancashire, Preston PR1 2HE, UK.
  • Department of Paediatric, East Lancashire Hospitals NHS Trust, Blackburn BB2 3HH, UK.

ORCID-Profile: Wei Quan

Förderung

  • University of Lancashire

Zitieren

APA

Han, B., Quan, W., Matuszewski, B., & Corbett, D. (2026). Respiratory Disease Classification Using NMF-Enhanced Log-Mel Spectrograms and Convolutional Recurrent Neural Networks. Sensors, 26(13), 4268. https://doi.org/10.3390/s26134268

Vancouver

Han B, Quan W, Matuszewski B, Corbett D. Respiratory Disease Classification Using NMF-Enhanced Log-Mel Spectrograms and Convolutional Recurrent Neural Networks. Sensors. 2026;26(13):4268. doi:10.3390/s26134268

BibTeX
@article{Han2026,
  author = {Han, Bowen and Quan, Wei and Matuszewski, Bogdan and Corbett, Dennis},
  title = {Respiratory Disease Classification Using NMF-Enhanced Log-Mel Spectrograms and Convolutional Recurrent Neural Networks},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {13},
  pages = {4268},
  doi = {10.3390/s26134268}
}
Kostenlos starten →