Research Research Projects

Machine Learning for Radiologic Biomarkers of Acute Neurological Injury

We aim to develop a high-precision automatic image classifier to identify neurological injury on brain CT. Our approach uses machine learning methods trained from large datasets of annotated CT images. We leverage unique data collections curated from local clinical and rehabilitation services to support developing novel 3D image feature parameterizations and suitable inference algorithms.

Research Leads: 

Derek Merck, Ph.D.
Jeffrey Brock, Ph.D.
A topological mapping of traumatic brain injury CT scans - red nodes are scans with significant brain bleeds, giving hope for automated mobile diagnostics and triage." Attribution: "K. Murthy, D. Merck, L. Merck and J. Brock using Ayasdi Mapper.