2024
Ofertado
- Línea de investigación:
- Health informatics
- Descripción:
Neurodegenerative pathologies such as Parkinson, Alzheimer and others are important determinants of our life quality during the last stages of our lives. In this context, early diagnosis and detection can improve the quality of life of the patients. Diagnosis is done using different nature information such as demographic information, motor and non-motor information extracted from clinical tests, bio-specimens and different types of scans. Among them, MRI has played a vital role in the characterization of multiple neurological diseases. When MRIs have been used with machine learning techniques for such an objective, most approaches include a previous thorough feature extraction process. One of the advantages of Deep Learning approaches is not needing such complex feature extraction processes. However, before the features are extracted, MRIs go through other processes such as registration, brain extraction, etc. The aim of this project would be to train a learner with different 3D images obtained in the initial preprocess stages of MRIs to asses to which extent are they effective in neurologic pathology detection.
Objectives:
- Selection of neural network architectures to work with MRI (3D images)
- Training models with different types of images for neurologic pathology detection
- Compare outcomes and conclude about the contribution of different processing stages
Titulaciones:
- Grado en Ingeniería Informática
- Computación
- Grado en Inteligencia Artificial
- Máster Universitario en Ingeniería Computacional y Sistemas Inteligentes
- Participantes:
-
- Director(es):
- Ibai Gurrutxaga
Javier Muguerza
Olatz Arbelaitz
- Universidad:
- Universidad del País Vasco/Euskal Herriko Unibertsitatea (UPV/EHU)
- Centro:
- Informatika Fakultatea - Facultad de Informática
- Departamento:
- Konputagailuen Arkitektura eta Teknologia - Arquitectura y Tecnología de computadores
- Año lectura:
- 2024