BUILDING AN APPLICATION TO IDENTIFY EXPIRATION DATES ON PACKAGING

Abstract

This paper presents the development of an expiry reminder mobile application designed to help users efficiently manage and track the expiration dates of their items. The application leverages image feature extraction and embedding techniques to identify and store expiration information in a vector database for accurate image-based searches. The frontend is built using Flutter, ensuring a seamless and intuitive user experience, while the backend, powered by FastAPI, provides robust and scalable functionality. PaddleOCR is integrated into the system for advanced text recognition, enhancing accuracy and performance in identifying expiration dates The app categorizes items into three expiry levels: expired, almost expired, and up to date, helping users prioritize their actions. A notification feature sends timely email reminders a set number of days before an item’s expiration date. The application also includes advanced search and filtering capabilities, enabling users to quickly retrieve relevant items based on expiration status or category. This application addresses the challenge of expiration management, promoting better resource utilization and reducing waste. Future improvements aim to incorporate AI-based text recognition for expiration date detection, expanding its capabilities to a broader range of item types and user needs.