UTILIZAÇÃO DE MÉTODO ENSEMBLE PARA CLASSIFICAÇÃO DE IMAGENS MÉDICAS CARDÍACAS:
Carregando...
Arquivos
Citações na Scopus
Tipo de produção
Trabalho de Conclusão de Curso
Data
2022-06-13
Autores
FELIPE MACIEL DE SOUSA
GUILHERME ORMOND SAMPAIO
LEON FERREIRA BELLINI
PEDRO FREITAS MAGALHÃES BARBOSA
GUILHERME ORMOND SAMPAIO
LEON FERREIRA BELLINI
PEDRO FREITAS MAGALHÃES BARBOSA
Orientador
Leila Cristina Carneiro Bergamasco
Periódico
Título da Revista
ISSN da Revista
Título de Volume
Citação
Texto completo (DOI)
Palavras-chave
Ressonância magnética cardíaca,cardiomiopatias,aprendizado de máquina supervisionado,método ensemble
Resumo
Doenças cardíacas figuram entre as principais causas de morte no mundo, segundo a
Organização Mundial da Saúde elas são responsáveis por cerca de 17 milhões de mortes globalmente.
Em resposta a esse cenário, as pessoas têm procurado com frequência acompanhamento
médico a fim de evitar um diagnóstico tardio. Esse processo promoveu um aumento
na demanda por exames de ressonância magnética cardíaca e a identificação de cardiomiopatias
nesse volume de exames representa um grande desafio para as equipes médicas. Embora
sistemas inteligentes sejam capazes de identificar doenças cardíacas nos exames mencionados
acima, poucas abordagens consideram diferentes slices do órgão e as especificidades do ciclo
cardíaco em sua análise. Este trabalho apresenta uma metodologia baseada em aprendizado de
máquina supervisionado para abordar essa análise considerando os aspectos citados. Durante
os experimentos foi registrada uma acurácia de 80,00% e precisão de 82,26% no melhor caso
de teste, que consiste na utilização das estruturas do epicárdio e endocárdio, durante o ciclo
diastólico e utilizando treze frames do ciclo
Heart diseases are among the most common causes of death in the world, according to the World Health Organization, said diseases are responsible for about 17 million deaths glob- ally. In response to this scenario, people have often sought medical attention in order to prevent late diagnosis. The aforementioned fact promoted an increase in demand for cardiac magnetic resonance exams. The identification of cardiomyopathies in this volume of exams represents a huge challenge for medical teams. Although smart systems are capable of identifying car- diac diseases in the exams mentioned above, few approaches consider different slices of the organ and the specificities of the cardiac cycle in its analysis. This essay presents a methodol- ogy based on supervised machine learning to address this analysis, considering the mentioned aspects. During the experiments, an accuracy of 80.00% and a precision of 82.26% were ob- tained in the best test scenario, which consisted of using the structures of the epicardium and endocardium during the diastolic cycle and utilizing thirteen cardiac cycle frames.
Heart diseases are among the most common causes of death in the world, according to the World Health Organization, said diseases are responsible for about 17 million deaths glob- ally. In response to this scenario, people have often sought medical attention in order to prevent late diagnosis. The aforementioned fact promoted an increase in demand for cardiac magnetic resonance exams. The identification of cardiomyopathies in this volume of exams represents a huge challenge for medical teams. Although smart systems are capable of identifying car- diac diseases in the exams mentioned above, few approaches consider different slices of the organ and the specificities of the cardiac cycle in its analysis. This essay presents a methodol- ogy based on supervised machine learning to address this analysis, considering the mentioned aspects. During the experiments, an accuracy of 80.00% and a precision of 82.26% were ob- tained in the best test scenario, which consisted of using the structures of the epicardium and endocardium during the diastolic cycle and utilizing thirteen cardiac cycle frames.