D-Cov19Net: A DNN based COVID-19 detection system using lung sound.

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Medicine
• Respiratory Medicine

Background:

• The COVID-19 pandemic highlighted the need for rapid and accessible diagnostic tools.
• Limitations in traditional testing methods necessitate innovative approaches for widespread screening.
• The potential of physiological signals, like respiratory sounds, for disease detection is increasingly recognized.

Purpose of the Study:

• To develop an automated system for COVID-19 detection using a single input parameter.
• To leverage deep learning for accurate and efficient diagnosis of respiratory infections.
• To create a user-friendly and feasible tool for early COVID-19 identification.

Main Methods:

• A Deep Convolution Neural Network (D-Cov19Net) was designed for audio signal analysis.
• The model was trained on a dataset of 23,592 respiratory sound recordings.
• Performance was evaluated using standard metrics like Area Under the Curve (AUC) and sensitivity.

Main Results:

• D-Cov19Net achieved a high Area Under the Curve (AUC) of 0.972.
• The model demonstrated excellent sensitivity, reaching 0.983 after 100 training epochs.
• The system proved effective in distinguishing COVID-19 positive cases from others based on lung sounds.

Conclusions:

• D-Cov19Net offers a highly accurate and sensitive method for COVID-19 auto-diagnosis.
• The system's simplicity, feasibility, and speed make it valuable for biomedical technology applications.
• This approach supports remote patient monitoring and social distancing during public health crises.