Currently working on the automated segmentation of mouse brain MRI using deep neural networks.
PhD course in Morphogenesis and Tissue Engineering, La Sapienza November 2018–Ongoing, in cotutela with University of Eastern Finland, Faculty of Health Sciences (Finlandia).
Deep Learning methods in MRI, under the tutoring of professor Federico Giove
Centro Fermi Rome: Research fellowship October 2018–Ongoing
Development of segmentation and multi-parametric analysis methods of rodent MRI images for the quantification of microstructural parameters, as part of the H2020 project T-MENS
International secondment Kuopio
Charles River, Centro Fermi January 2019–November 2019
Automated segmentation of mouse brain MRI with deep learning algorithms, scientific supervisor: professor Jussi Tohka
Institute for Complex Systems (ISC) Rome
In vivo MRS, Sapienza University July 2018–September 2018
Metabolite quantification with in vivo magnetic resonance spectroscopy data (MRS) as part of a study
on Huntington’s disease.
Institute for Complex Systems (ISC) Rome
Post-graduate Internship, Sapienza University April 2017–December 2017
Pre- and post processing of human brain NMR images acquired with diffusion MRI protocols.
Sintesi progetto di ricerca
Image segmentation is a common step in the analysis of pre-clinical brain MRI, often performed manually. An alternative to manual segmentation is automated, registration-based segmentation, which suffers from a bias owed to the limited capacity of registration to adapt to pathological conditions. In this work a novel method is developed for the segmentation of small rodent brain MRI based on Convolutional Neural Networks (CNNs). The experiments presented show how CNNs provide a fast, robust and accurate alternative to both manual and registration-based methods.
Using the segmentation masks thus generated I then extracted 39 parameters characterizing the position and orientation of the hippocampus in rats 5 months after traumatic brain injury, allowing for the discrimination between epileptic and non-epileptic animals on a purely anatomical basis, with a balanced accuracy of 0.8.