xRayAID Detecting Pneumonia Using Artificial Intelligence

Abstract

Pneumonia is a type of acute respiratory infection that impacts people's lives in several ways, demanding an accurate and fast diagnosis. High death rates, massive socioeconomic impacts, and a significant gap between the number of available doctors based on its geographic location are some of the problems surrounding this topic. The xRayAID is a tool that uses machine learning to assist doctors in diagnosis of pneumonia on frontal chest radiographs. That was done by using a modified DenseNet-121 neural network architecture trained on the Radiological Society of North America (RSNA) public dataset. The results showed that this tool is able to help doctors to identify pneumonia scenarios, achieving a validation accuracy of 87.9%.