Calculating Insulin With Automated Carb Counting Using AI, ML and Web Bluetooth
Presented by: Todd Sharp
Diabetes is a disease that affects 400 million people worldwide. Managing the disease means eating a healthy diet, getting routine exercise, monitoring blood sugar levels and administering insulin to manage those levels. The good news is that technology exists to help people manage the disease. Continuous Glucose Monitors, Insulin pumps and even open source "artificial pancreas" solutions exist and help diabetics to manage their blood sugar levels with very little time and effort. All of these technologies make life easier, but they all still require the user to manually calculate and enter their carbohydrate consumption in order to calculate the necessary insulin dose.
My 13-year-old daughter was recently diagnosed with Type 1 Diabetes and we have been slowly learning how to help her manage the disease. I was inspired to try to help solve the problem of carbohydrate counting so I came up with a Progressive Web Application that can be used to automate that process using things like image recognition, machine learning, web bluetooth, microservices and more. In this session we'll look at some of the technologies behind the application and how they all work together to try and solve the difficult problem of counting carbohydrates in order to calculate a necessary insulin dose.