Medical image analysis is hindered by models built from limited data

No quantitative or standard way to measure disease severity

Existing active learning systems lack multi-modal data annotation capability