ResNet-EfficientNet powered framework for high-precision cough-based classification of infectious diseases

Area of Science:

• * Medical Diagnostics
• * Artificial Intelligence in Healthcare
• * Respiratory Illnesses

Background:

• * COVID-19, caused by SARS-CoV-2, is a highly contagious respiratory illness that emerged in late 2019, causing global health and economic challenges.
• * Rapid diagnosis of COVID-19 is crucial for controlling transmission and managing patient care, similar to other infectious diseases.
• * Transmission occurs via respiratory droplets, highlighting the need for efficient diagnostic tools.

Purpose of the Study:

• * To propose a deep learning framework for classifying COVID-19 using cough sounds.
• * To evaluate the performance of various deep learning models, including 1D-CNN, DS-CNN, EfficientNet v2, and ResNet, for COVID-19 detection via cough analysis.
• * To identify the most effective model for accurate and rapid COVID-19 diagnosis based on cough acoustics.

Main Methods:

• * Development of a deep learning framework for analyzing cough sounds.
• * Implementation and comparison of multiple deep learning architectures: 1D-CNN, DS-CNN, EfficientNet v2, and ResNet.
• * Performance evaluation using key metrics: accuracy, precision, recall, F1-Score, Matthews Correlation Coefficient (MCC), and False Positive Rate (FPR).

Main Results:

• * Pre-trained models, specifically EfficientNet v2 and ResNet, demonstrated superior performance compared to other deep learning models.
• * The ResNet model achieved exceptional results with 98.5% accuracy, 98.99% precision, 98% recall, a 0.9849 F1-Score, a 0.9699 MCC, and a 0.01 FPR.
• * These findings indicate ResNet's superiority in classifying COVID-19 positive cough sounds.

Conclusions:

• * The proposed deep learning framework, particularly using the ResNet model, offers a highly accurate method for COVID-19 classification via cough sounds.
• * This approach facilitates early intervention, enabling physicians to isolate or treat patients promptly and reduce disease transmission.
• * Cough sound analysis using advanced AI presents a promising, non-invasive tool for widespread COVID-19 screening and management.