Projects
PhD Project
The aim of my PhD is to implement deep learning algorithms to medical images of glioma to both stage the treatment of this disease and generate synthetic MRI data.
Fixel Based Analysis
This project aims at finding biomarkers for migraine using the fixel-based analysis pipeline of MRtrix3 and diffusion-weighted magnetic resonance imaging data.
Master Thesis
My master thesis, titled "Investigating Structural Connectivity in Episodic Migraine using Graph Theory", searched for biomarkers of episodic migraine using DWI data. Tractography methods were applied and a connectome was made, from which connectivity metrics were analysed and compared.
Boat Management System
This project was built in the scope of the course "Information Systems and Databases" and it consisted in the development of a database model (E-A model), of writing SQL queries and OLAP queries, integrity constraints, and the creation of a web application prototype.
Machine Learning Project
This project entailed the resolution of several machine learning problems not only of regression but also of classification. The following algorithms were used: Linear/Lasso/Ridge Regression, Linear Support Vector Regression, Stochastic Gradient Descent, Gaussian Processes, Convolutional Neural Networks, Multilayer Perceptron, Support Vector Machine, Random Forest and k-Nearest Neighbours. Additionally, k-fold cross validation, outlier detection and ensemble learning were also used.
Sleep Stage Classification
This project was made in the scope of the course "Signal Processing in Bioengineering" and aimed at the classification of sleep stages from multi-modal Polysomnogram (PSG) recordings. Preprocessing techniques were evaluated, namely ICA vs conventional filtering, and a support vector machines algorithm was used for the classification. The final minipaper can be found on the LaSEEB's website.
Heart Rate and Pulse Transit Time Sensor using PPG and ECG
This project's main goal was to develop a sensor that could calculate the heart rate and pulse transit time in real time. To that end, an Arduino Uno was used as well as ECG and PPG sensors. Real time processing (using Arduino) and post processing (using MATLAB) approaches were compared.
Cancer Detection Algorithm
This project's main goal was to develop a classifier based on Bayes Networks that could predict if a person has cancer given certain biomedical data. This project was done in Java, where two applications were created: one to read the data and to learn from it and another to, given a set of data, predict, with a certain probability, if that person had cancer.