Comparison of AI Models for Data Extraction in Glucometers

Diabetes is a chronic condition that requires continuous monitoring of blood glucose levels, making glucose meters essential tools for managing patients’ health. This study compares three Artificial Intelligence models — Gemini, GPT-4o, and Llava 1.5 — to identify which one extracts glucose, date, and time data from these devices’ images with greater accuracy and efficiency. Through prompt engineering techniques, the aim is to optimize the extraction process, making it more reliable and automated, contributing to more precise and accessible glucose monitoring.