Malte Hoffmann


Instructor in Radiology, Harvard Medical School

Research Staff, Massachusetts General Hospital

Affiliated Staff, Boston Children’s Hospital

Athinoula A Martinos Center for Biomedical Imaging

149 13th St, Suite 2301

Charlestown, MA 02129

United States

mhoffmann at mgh dot harvard dot edu

Malte's photo

Research interests

Subject motion is a major confound in magnetic resonance imaging (MRI), as it is almost impossible to avoid artifacts if subjects do not remain still for the length of a scan; my research focuses on the development of techniques for detecting and correcting motion in the scanner. In particular, I am interested in methods using image registration, which capture the spatial relationship between objects directly from images.

Recently, my work builds on innovations in deep learning, which have enabled algorithms with unprecedented speed and reliability. My contributions include a strategy for learning registration from noise distributions, eliminating the need for acquired training data while providing deep neural networks with the previously missing ability to generalize to unseen data types.

My current goal is to automate and enhance the motion resilience of fetal MRI, which is challenging due to excessive levels of motion. Exploiting advances in deep learning and integrating these into the acquisition pipeline will unlock the full potential of the modality.

Published work

Peer-reviewed publications

Peer-reviewed conference abstracts


Honors and awards