Regresar

A Convolutional Neural Network-Based Web Prototype to Support COVID-19 Detection Using Chest X-rays

Abstract:

COVID-19 continues to cause health problems to humanity. Some fields of science have conducted research to mitigate and reduce the harmful effects of this virus. In the healthcare field, radiographs are very important because they provide data that allow detection and assessment of pathologies in a reliable way. In this context, machine learning and data mining provide the mechanisms and algorithms that can support health care activities. Machine learning capability allows the neural network to learn, identify and interpret the results of a radiographs set. With these considerations, this research develops a web prototype based on convolutional neural networks to support the detection of COVID-19 using chest X-rays. For this, two sequential phases were defined, namely: data mining and software development. In this context, Cross Industry Standard Process for Data Mining (CRISP-DM) was used to select the deep convolutional neural network that best fits our case study. With this previous analysis, a web prototype was developed using two frameworks: Flask (for backend) and Angular (for frontend). Conclusions and future work are described at the end of the document.