Computing Degree Show 2018

Automated Brain Tumour Segmentation based on MR Images using U-Net and its application for Mutation Status Prediction

Machine Learning can be used for predicting the mutation status of brain tumours. If accurate enough, this can provide enormous benefits for healthcare, where much less time would have to be spent on analysing brain tumours, it could even make it unnecessary to perform unpleasant biopsies on patients. Such a solution was built in-house and was used on data which was annotated by medical experts. This project focuses on automatically annotating medical data rather than by the hand of an expert and then evaluating the prediction performance.

Philipp Schnell (paschnell@dundee.ac.uk)