AI Based Food Freshness Tracking System with Shelf-Life Prediction

To predict food spoilage before visible signs of spoilage, this paper introduces the design, development, and testing of an artificial intelligence (AI) based food freshness monitoring system that combines machine learning, environmental sensing, and gas sensing. Static expiration dates form the foundation of traditional food safety practice, which has a tendency to result in eating spoiled food or premature disposal of fresh food. Unlike binary classification, our system uses regression-based machine learning models combined with multiple sensors to provide precise shelf-life estimation. Temperature, humidity, and gas emissions are three major spoilage indicators that can be monitored by a prototype combined on an Arduino board. The system can also predict shelf-life estimates of various types of foods with more than $90 \%$ accuracy. Unlike old-fashioned expiration, user testing demonstrated a significant reduction in self-reported food waste. This work is a contribution to the new field of AI-based food safety monitoring and has implications for reducing food waste in home and commercial settings.