Michael Roberts

I research and develop image analysis techniques and apply them to problems in medical imaging -- principally variational and machine learning based methods. In particular, I look to detect imaging or clinical biomarkers indicating disease progression. My background is mathematics and image analysis. Developing new and robust image segmentation methods was my primary research area for my Ph.D. and developing new deep neural networks for medical image segmentation was the focus of my previous postdoc. I lead the algorithm development for the AIX-COVNET COVID-19 collaboration in Cambridge, developing machine learning-based methods for COVID-19 prognosis and diagnosis using both imaging and non-imaging data.