Artificial intelligence meets diabetes at the micro-scale: forecasting complications and quantitatively mining insulin granules motions

Diabetes is a chronic, metabolic disease characterized by elevated levels of blood glucose, which leads over time to serious complications. Unfortunately, no cure exists and the only way to tackle the disease consists in constantly monitoring the blood sugar concentration and furnishing insulin via periodic injections in type 1 diabetes and only in prevention in type 2. The thesis presents innovative applications of Artificial Intelligence to two key issues in diabetes. In a first part, it presents a decision support system that analyses red blood cells fluidity to diagnose diabetes and diabetes with complications. In the second part, it investigates the insulin granules’ dynamic inside the b-cells, assessing their motion in presence of different environmental glucose concentration with a clustering analysis.