Unhealthy dietary patterns and excessive intake of processed foods are major contributors to the global rise of obesity and chronic diseases, highlighting the need for accessible tools that enable consumers to make informed food choices at the point of purchase. Label-AI is a web-based system designed to address this challenge by scanning product barcodes by scanning a product’s Universal Product Code (UPC) with a smartphone, LabelAI retrieves detailed nutrient data from an extensive food database (Open Food Facts) extracting nutrition information and generating a NutriScore-style health rating on a scale of 0-10. The system’s engine processes the nutritional information obtained from barcode scans and computes rating on a 0–10 scale based on key nutrients such as sugars, fat, saturated fat, salt, proteins, fiber, and energy per 100 g. Products with lower scores trigger alerts and suggestions for healthier alternatives within the same category. This paper presents the design and evaluation of Label-AI, including an overview of existing barcode-based nutrition applications, a two-tier architecture that combines browser-side scanning with cloud-based data retrieval, and a hybrid scoring mechanism that integrates machine learning with rule-based thresholds.

Barcode scanning; Nutrition tracking; Mobile health (mHealth); Personalized diet; Artificial intelligence; Machine Learning; Rule-based Scoring; Open Food Facts