3D medical image segmentation
Encadrants : Isabelle Bloch (LIP6), Sarah Stricker and Giammarco La Barbera (Hôpital Necker, sarah.stricker@gmail.com, giammarco.labarbera93@gmail.com)
Disponible : OUI
Spécialité : IMA
Nombre d'étudiants : 1
Description :
Description: Post-hemorrhagic hydrocephalus is a medical complication prevalent in premature infants, characterized by abnormal cerebrospinal fluid (CSF) accumulation and hence dilation of the ventricles. This can be assessed using magnetic resonance imaging (MRI) of the brain. The objective of this project is to achieve automatic segmentation of the ventricles in pre and post-operative MRIs of premature infants.
Pré-requis : Image analysis and machine learning
Travail demandé : 1. Literature review
2. Development of 3D segmentation methods based on deep learning. Since data may not be numerous enough, data augmentation methods adapted to the problem will be designed, as well as training from synthetic data. Based on the segmentation results, the volume of the ventricles can be computed.
2. Development of 3D segmentation methods based on deep learning. Since data may not be numerous enough, data augmentation methods adapted to the problem will be designed, as well as training from synthetic data. Based on the segmentation results, the volume of the ventricles can be computed.
