AUTOMATED LUNG DISEASE DETECTION USING HANDCRAFTED ACOUSTIC FEATURES DERIVED FROM RESPIRATORY SOUNDS
International Journal of Progressive Research in Engineering Management and Science · 01.07.2026
DOI: 10.58257/ijprems54441 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Progressive Research in Engineering Management and Science, online erschienen am 01.07.2026.
Abstract (Originalsprache)
Respiratory diseases remain a major global health concern, requiring timely and accurate diagnosis to improve patient outcomes.This study presents an automated lung disease detection framework based on respiratory audio recordings using machine learning and deep learning techniques.The publicly available ICBHI Respiratory Sound Database was utilized, and the dataset was restructured by removing Asthma and Lower Respiratory Tract Infection (LRTI) classes while merging Bronchiectasis and Bronchiolitis into a single Bronchial_Disease category, resulting in a five-class classification problem with 915 audio samples.Acoustic features, including Mel-Frequency Cepstral Coefficients (MFCC), Chroma Short-Time Fourier Transform (Chroma STFT), Mel Spectrogram, Spectral Contrast, and Tonnetz, were extracted to effectively represent respiratory sounds.To address class imbalance, mild Synthetic Minority Oversampling Technique (SMOTE) and class weighting were employed during model training.The extracted handcrafted acoustic features were used to train and evaluate an Artificial Neural Network (ANN) and a onedimensional Convolutional Neural Network (1D-CNN).Experimental results demonstrated that the proposed framework achieved high classification performance.Among the evaluated models, the ANN achieved superior overall performance, attaining an accuracy of 91.26%, while exhibiting better training stability and generalization than the 1D-CNN on the extracted handcrafted features.Furthermore, the ANN model, identified as the best-performing classifier, was integrated into a Flask-based web application to facilitate automated respiratory disease prediction from lung sound recordings.
Zitieren
APA
(2026). AUTOMATED LUNG DISEASE DETECTION USING HANDCRAFTED ACOUSTIC FEATURES DERIVED FROM RESPIRATORY SOUNDS. International Journal of Progressive Research in Engineering Management and Science. https://doi.org/10.58257/ijprems54441
Vancouver
AUTOMATED LUNG DISEASE DETECTION USING HANDCRAFTED ACOUSTIC FEATURES DERIVED FROM RESPIRATORY SOUNDS. International Journal of Progressive Research in Engineering Management and Science. 2026. doi:10.58257/ijprems54441
BibTeX
@article{euvimed2026,
title = {AUTOMATED LUNG DISEASE DETECTION USING HANDCRAFTED ACOUSTIC FEATURES DERIVED FROM RESPIRATORY SOUNDS},
journal = {International Journal of Progressive Research in Engineering Management and Science},
year = {2026},
doi = {10.58257/ijprems54441}
}
EUVIMED