Marcelo Pereyra

Heriot-Watt University


On convexity, regularisation, and Bayesian estimation in imaging inverse problems


Statistics Seminar


28th February 2020, 2:00 pm – 2:45 pm
Fry Building, G. 13


This talk presents an overview of new theory regarding the role of convexity in Bayesian estimation in high-dimensional inverse problems. We pay special attention to connections between convexity, MAP and MMSE point estimation, and hypothesis testing. This new theory is illustrated with examples from imaging sciences related to astronomical and medical imaging.





Organiser: Juliette Unwin

Comments are closed.
css.php