LWCNN-Based Deep Learning Model for COVID-19 Detection: Performance Analysis on Voice Data

Rapid and accurate detection of COVID-19 is critical in combating the pandemic. In this study, a deep learning-based model is proposed to classify COVID-19 patients as positive, symptomatic, and asymptomatic. The model utilizes features extracted from voice data and is compared with different convolutional neural network (CNN) architectures for COVID-19 detection. Experiments conducted with commonly used networks such as AlexNet, GoogleNet, ShuffleNet, and SqueezeNet demonstrated that the proposed model outperforms them in terms of overall accuracy, precision, and F1-score. The model’s ability to maintain a better balance between classes and minimize misclassification errors suggests its potential for clinical applications. The findings indicate that the developed model provides a reliable and effective alternative for COVID-19 detection. Future work aims to test the model on larger datasets and support it with different biomedical signals.