Fighting Diabetes With Technology
summary
A working proof‑of‑concept Progressive Web App automates carb counting by accepting photo/voice/typed inputs, calling a third‑party image recognition service (Clarifai), using Microsoft OCR, querying USDA FoodData Central for carbs, and pairing via Web Bluetooth to a scale and glucose meter. Backend is three microservices (Java/Node/Micronaut) on Kubernetes with Oracle Autonomous DB (ORDS/SODA). Demo reads a 197 g apple, totals ~79 g carbs and computes ~8.65 units (rounded to 9); model uses Oracle ML linear regression to predict next glucose.
AI-written summary · about
chapters (11)
- 0:00Start
- 11:14What's Missing?
- 23:18Cloud event triggers serverless function Fn uses image recognition to identify concepts and OCR to recognize text
- 25:25Concept with highest confidence used - Search USDA Food Data Central API
- 26:47Serverless Function: Image Recognition & OCR
- 29:35Front End: Installable PWA with Web Bluetooth
- 33:21User Service: Helidon SE with ORDS
- 33:37Oracle REST Data Services gets your data in and out of Oracle DB without SQL queries using the REST protocol.
- 41:50Formula Service: Node.JS with SODA
- 47:08Nutrition Service: Micronaut with Micronaut Data (JDBC)
- 52:38Summary
related talks
09mLaunch by Veri
27mmulti-modal AI to avoid eating takeaways
13mIntelligent Health AI Conference September 2020 - Diabetes Digital Media (DDM)
23mShip It First, Fix It Later: How a Medical Crisis Led to a Global Tool
17mTransforming access to healthcare through Web AI
23mConnecting Food Journeys through AI and the Open Food Platform
30mUnleashing AI in Your Web Apps
33mKeynote: Air Fryers, Automation, and AI