ATwC2N2: Automatic Chronic Obstructive Pulmonary Disease Classification Using Adaptive Two-Way Cascade Convolutional Neural Network
Journal of Innovative Image Processing · 23.07.2026 · 8 (3) · 1196-1220
DOI: 10.36548/jiip.2026.3.021 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Innovative Image Processing, online erschienen am 23.07.2026. Verfasst von Nagendra Kumar V V. und 1 weiteren Autor:innen. Die Arbeit nennt 23 Literaturangaben.
Themen: COPD
Abstract (Originalsprache)
Chronic Obstructive Pulmonary Disease (COPD) is a widespread lung condition that impacts the quality of life and lung functioning, leading to significant morbidity and mortality on a global scale. Non-invasive lung sound analysis is a method that has garnered growing interest in detecting COPD at an early stage and seeking immediate medical assistance. However, manually assessing respiratory sounds to determine the severity of COPD is subjective and prone to human error, often resulting in inaccurate assessments. This highlights the need for automated and accurate monitoring systems. This paper proposes an Adaptive Two-Way Cascade Convolutional Neural Network (ATwC2N2) optimized by the Adaptive Snake Swarm Optimization (AS2O) algorithm for classifying COPD stages using respiratory sounds. The process involves cleaning and normalizing the respiratory sound, transforming it into a spectrogram for feature extraction and disease classification. The model utilizes a dual-pathway feature extraction mechanism that processes raw lung sound signals using a 1D CNN and spectrogram images using a 2D CNN simultaneously, integrating temporal and frequency domain features for superior classification performance. The AS2O algorithm is used to optimize network hyperparameters, such as the learning rate, batch size, convolutional filters, dropout rate, and fully connected layer size, thereby enhancing feature learning, convergence stability, and overall classification performance. The proposed system categorizes subjects into four groups: healthy, mild COPD, moderate COPD, and severe COPD. The ATwC2N2 model achieves an accuracy of 97.14%, specificity of 96.01%, precision of 95.91%, recall of 98.32%, and an F1-score of 97.10%. The proposed model improves overall accuracy by 0.88%, 67.82%, 2.06%, and 3.07% compared to EMD-IMF-ellipse area from 2D-PSR, STFT + Wavelet, Expert Diagnostic System, and ALSD-Net, respectively.
Zitieren
APA
V., N. K. V., & L., K. (2026). ATwC2N2: Automatic Chronic Obstructive Pulmonary Disease Classification Using Adaptive Two-Way Cascade Convolutional Neural Network. Journal of Innovative Image Processing, 8(3), 1196-1220. https://doi.org/10.36548/jiip.2026.3.021
Vancouver
V. NKV, L. K. ATwC2N2: Automatic Chronic Obstructive Pulmonary Disease Classification Using Adaptive Two-Way Cascade Convolutional Neural Network. Journal of Innovative Image Processing. 2026;8(3):1196-1220. doi:10.36548/jiip.2026.3.021
BibTeX
@article{V2026,
author = {V., Nagendra Kumar V and L., Kavisankar},
title = {ATwC2N2: Automatic Chronic Obstructive Pulmonary Disease Classification Using Adaptive Two-Way Cascade Convolutional Neural Network},
journal = {Journal of Innovative Image Processing},
year = {2026},
volume = {8},
number = {3},
pages = {1196--1220},
doi = {10.36548/jiip.2026.3.021}
}
EUVIMED