Machine Learning Techniques for Disease Localisation in Medical Imaging

In this presentation, we will focus on a specific class of machine learning techniques that aim at the ‘detection of rare events', which is known as ‘anomaly detection.’ After a brief introduction to the problem and the methods, I will present preliminary results of their application to prostate cancer localisation using processed ultrasound contrast data. The challenges of the approach will be discussed as well as possible future research directions.