Machine Learning for Brain Tumour Detection: Comparing Probabilistic, Classical, and Deep Learning Models

Machine Learning for Brain Tumour Detection: Comparing Probabilistic, Classical, and Deep Learning Models

Statistics

Ignacio Vigil Cardenal

2005

This research project aims to comprehensively implement and evaluate several machine learning models for the critical task of brain tumour classification using Magnetic Resonance Imaging (MRI) data. A classical Support Vector Machine (SVM) will serve as a baseline, against which the performance of advanced deep learning architectures, including Convolutional Neural Networks (CNNs) and a Bayesian Convolutional Neural Network (BCNN), will be assessed. The study will rigorously compare these models across key performance indicators: classification accuracy, the robustness of uncertainty quantification, and computational efficiency. Ultimately, this work seeks to provide valuable insights into the efficacy and utility of diverse machine learning paradigms in medical image analysis, with a particular focus on the crucial role of uncertainty estimation in supporting more reliable and informed decisions for early brain tumour diagnosis.

Mathematics, Machine Learning, Statistics, Computer Vision, Deep Learning, Bayesian, Medical Imaging