Detecção de anomalias médicas no pulmão usando técnicas de inteligência artificial

• Autor(a) principal: Silva, Rafael Reinaldo
• Data de Publicação: 2024
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional do IFAM (Repositório Institucional do Instituto Federal do Amazonas)
• Texto Completo: http://repositorio.ifam.edu.br/jspui/handle/4321/1403
• topic: Inteligência artificial; Redes Neurais; Pneumonia; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA::ELETRONICA INDUSTRIAL, SISTEMAS E CONTROLES ELETRONICOS::CONTROLE DE PROCESSOS ELETRONICOS, RETROALIMENTACAO

Artificial Intelligence (AI) is not a contemporary concept, and its presence is not alien to society; currently, numerous applications employ this knowledge. However, despite the growing incorporation of technology in various contexts and increasingly in everyday life, a significant portion of the population still views technological intervention with reservations. In the medical field, it is no different; there is concern about the replacement of healthcare professionals by intelligent machines. Those who think so are mistaken; currently, AI is already involved in various aspects, such as disease prognosis, image understanding, interconnection of prescription databases, among others, and yet there is still a need for human presence. However, such advances do not spread uniformly; there will still be many gaps in the democratization of the application of this resource, whether due to a lack of technology, training, or even distances. In the latter case, it is more common in Brazil, which has continental extensions. According to the World Health Organization (WHO), in the last decade, two of the four leading causes of death in the world are related to respiratory diseases. In Brazil, according to the Ministry of Health, in the first nine months of 2022, there were more than 40,000 deaths from pneumonia alone. This work aims to compare machine learning algorithm models to present the pros and cons of each. Neural networks were used to check a dataset of images for training, validation, and testing of models with the purpose of recognizing and diagnosing cases of pneumonia through only X-ray examinations.