Lw-CovidNet: Lightweight deep-learning module for disease prediction in CT scan images

Covid-19 had intensely impacted the global health, which had spread illness and mortality worldwide. Throuhout this crisis, potent detection methods are vital for timely treatment and diagnosis. In this study, we introduce Lw-CovidNET, an acutely lightweight Depthwise Separable Convolutional Neural Network (DS-CNN) helping in the analysis and detection of COVID-19 in medical images. Our Lw-CovidNET model fused DSCNN layers, which offer precedence in terms of parameter effectiveness and computational cost compared to conventional CNNs. Lw-CovidNET comprises approximately 9M parameters, which records for its shortness and lightweight attributes. Empirical assessment done on SARS-CoV-2 CT-scan dataset signify strong results, achieving nearly 92% accuracy, 94% recall, 91% precision, and F1-score of 93%, signifying its potential for better diagnostic accuracy in medical imaging applications