Publication: Aplicación de deep-learning en neuroimágenes para el diagnóstico de enfermedades neurodegenerativas
Loading...
Official URL
Full text at PDC
Publication Date
2022
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
El Alzhéimer es la principal causa de demencia en el mundo, típicamente caracterizado por la pérdida de memoria seguido por una decline cognitivo. Gracias a avances recientes en el campo de la neuroimagen, como las Resonancias magnéticas (MRI) o la tomografía de emisión de positrones (PET) para el diagnóstico de esta enfermedad, se han propuesto diferentes técnicas dentro del campo del machine learning para dar lugar a soluciones y diagnósticos. Aquí hemos desarrollado una aplicación basada en aprendizaje profundo para identificar patrones de características comunes entre las neuroimágenes de pacientes diagnosticados con Alzhéimer, este tipo de herramientas tienen una naturaleza de caja negra al tratarse de tecnologías muy complejas. Esto abre la puerta al segundo objetivo del proyecto, usar varios algoritmos y técnicas de interpretación para explicar el diagnostico obtenido de nuestros modelos mediante mapas de calor.
Alzheimer’s is the leading cause of dementia in the world [1], typically characterized by memory loss followed by cognitive decline. Thanks to recent advances in the field of neu roimaging, such as Magnetic Resonance Imaging (MRI) or positron emission tomography (PET) for the diagnosis of this disease, different techniques have been proposed within the field of machine learning to give rise to solutions and diagnoses. Here we have develo ped an application based on deep learning to identify patterns of common characteristics in the neuroimaging of patients diagnosed with Alzheimer’s. These types of tools have a black box nature as they are very complex technologies. This opens the door to the second objective of the project, the use of various algorithms and interpretation techniques to explain the diagnosis obtained from our models through heat maps.
Alzheimer’s is the leading cause of dementia in the world [1], typically characterized by memory loss followed by cognitive decline. Thanks to recent advances in the field of neu roimaging, such as Magnetic Resonance Imaging (MRI) or positron emission tomography (PET) for the diagnosis of this disease, different techniques have been proposed within the field of machine learning to give rise to solutions and diagnoses. Here we have develo ped an application based on deep learning to identify patterns of common characteristics in the neuroimaging of patients diagnosed with Alzheimer’s. These types of tools have a black box nature as they are very complex technologies. This opens the door to the second objective of the project, the use of various algorithms and interpretation techniques to explain the diagnosis obtained from our models through heat maps.
Description
Trabajo de Fin de Grado en Ingeniería Informática, Facultad de Informática UCM, Departamento de Arquitectura de Computadores y Automática, Curso 2021/2022.