Effect of decay in magnetic resonance imaging on deep neural networks

• Autor(a) principal: Prazeres, Carlos Leandro Silva dos
• Data de Publicação: 2022
• Outros Autores: Paula, Perseu Lúcio Alexander Helene de, Monte, Mozart Nolasco, Estácio , Marcela Costa Alcântara, Santos, Esdras Adriano Barbosa dos, Campos, Laelia
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Research, Society and Development
• Texto Completo: https://rsdjournal.org/index.php/rsd/article/view/31868
• topic: Magnetic Resonance; Brain Tumors; Deep learning.; Resonancia magnética; Tumores del Cerebro; Aprendizaje profundo.; Imageamento por Ressonância Magnética; Tumores Cerebrais; Aprendizado profundo.

In the last decades, tasks of classification and segmentation of clinical findings using convolutional neural networkshave grown significantly in the sphere of radiology, and more precisely in the modality of magnetic resonance imaging. However, little is known about the behavior of the proposed architectures when faced with factors that degrade the spatial resolution and contrast resolution, since most models are trained with high quality images, which is not consistent with the general daily life. Therefore, it is necessary to analyze the performance of pre-trained neural networks under conditions in which there is deterioration of the input image. In this work, the effects of degradation of the resolutions were evaluated, both in classification and segmentation tasks of brain tumors, for three architectures: Mobilenet, Vgg16 and SEResNeXt50. The results obtained showed that the tasks performed are greatly affected by image quality distortions, especially in cases where the deteriorations become more intense.