Multi-Scale Convolutional Neural Network with Attention Mechanism for COVID-19 Detection from CT Images

Abstract

The COVID-19 pandemic has spurred urgent developments in medical imaging technologies for accurate and timely diagnosis. Computed Tomography (CT) imaging, in particular, offers detailed insights into lung abnormalities associated with COVID-19 infections. This paper introduces a novel Multi-Scale Convolutional Neural Network with Attention Mechanism (MS-CNN-Att) designed to enhance COVID-19 detection accuracy from CT images. The model integrates multi-scale feature extraction capabilities with attention mechanisms to highlight relevant patterns indicative of COVID-19 pathology. Experimental results demonstrate the efficacy of the proposed approach in achieving high diagnostic accuracy, thereby supporting clinical decision-making and public health management strategies during the ongoing pandemic.