Utilizando transfer learning para apoio a tomada de decisão na classificação de pneumonia usando imagem de raio-x

Abstract

This work explores the use of deep learning techniques, such as Convolutional Neural Networks (CNNs) and the YOLO architecture, for the detection of lung diseases through chest X-ray images acquired via IoT devices. A balanced dataset of 3,252 images was used for training, and the adaptation of YOLO for classification tasks showed promising results in identifying pneumonia and tuberculosis. The best performance was observed in the fifth epoch, emphasizing the system’s potential as a decision-support tool in regions with limited access to medical specialists. The integration of these technologies can significantly expand access to accurate and efficient diagnostics in public health contexts.