Phenotyping neurodegeneration using machine learning:

We build machine learning frameworks to process multimodal data and identify specific signatures of neurodegeneration. We have experience in dealing with large data cohorts such as the Framingham Heart Study and established several computational pipelines to efficiently process volumetric images of the brain, neuropathology and other modes of data and use them for further analysis. Click on Papers for full list.  

Digital pathology:

We develop computational frameworks based on deep learning to assist the pathologist. Our current application areas include kidney disease, lung cancer and colorectal cancer. Click on Papers for full list. 

Machine learning for musculoskeletal diseases:

We develop machine learning frameworks to bring efficiency to the analysis of large-scale studies such as Osteoarthritis Initiative (OAI) and Multicenter osteoarthritis (MOST) study. We are particularly interested to quantify structures that are responsible for pain and factors that contribute to the progression of knee osteoarthritis. Click on Papers for full list.