CT-FastNet: Detecção de COVID-19 a partir de Tomografias Computadorizadas (TC) de Tórax usando Inteligência Artificial

Abstract

Many countries have been affected by the COVID-19, and health departments are facing delays to detect the new coronavirus symptoms. Artificial Intelligence (AI) models are designed for the automatic detection of respiratory diseases patterns using computed tomography (CT) scans of the chest. However, the training time consumed by the algorithms is a key parameter that is not properly attended. In this article, we propose an AI solution using an activation function that helps to obtain a low training time. Experimental results show that our proposal overcome several deep learning architectures, such as the 3D deep Convolutional Neural Network to Detect COVID-19 (DeCoVNet).