Interdisciplinarity is what drives modern research…
(Note: this page is currently under development and will be filled with more details. In the meantime, see also reuter.mit.edu and our publications )
The lab is committed to publishing our code ( Deep-MI Github ) and posting manuscripts on preprint servers.
The research focus the DeepMI lab is on the development of a novel deep-learning (CNNs) and machine-learning methods for the automated analysis of medical images, such as human brain MRI. In addition to medical image computing and computational neuroimaging, our research intersets include computational geometry and topology, computer and biomedical vision, computer-aided design, geometric modeling and computer graphics.
In close collaboration with clinical and industrial partners, we develop the next-generation computational techniques for the analysis of large biomedical image datasets, including:
Development of innovative methods for reliable image acquisition, processing, analysis and interpretation – in particular sensitive longitudinal analysis, registration, reconstruction, segmentation and predictive modeling.
Deep- and machine-learning, computer vision and statistical modeling for the extraction of biomarkers and other clinically relevant information from large data sets (big data) – in particular for computer-aided diagnosis (at presymptomatic stages) and prognosis, personalized medicine, treatment planning, patient stratification, and identification of risk- or preserving factors of neurodegenerative disease.
Our research focuses on improving our understanding of brain development, neurodegeneration (aging, dementia, Alzheimer’s disease, Huntington’s disease), the sensitive quantification of subtle drug effects, and improving tumor treatment assessment, via multi-modal imaging.