Product Auto-Check Consumer Expiry (PACE): A Unified Deep Learning Framework for Shelf-Life Management

Read full text Summarise with VIRAT
Authors: Shuaib M, Jagadeesh K, Suraj Kumar B P, Mogesh K, Keerthika A
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Product expiration is a growing problem for global retailers and modern consumers. It leads to significant financial losses, legal penalties, and serious public health risks. In retail, expired inventory can completely wipe out cost price and lower profits. Unsold products can make up to 50% of an organization’s annual profit in some cases. Consumers, on the other hand, face the risk of illness or even death from consuming spoiled items or expired medications. Traditional manual tracking and standard Optical Character Recognition (OCR) tools like Tesseract are often inefficient and can make mistakes. They frequently misread industrial dot-matrix fonts or fail completely with rotated samples. This paper proposes PACE (Product Auto-Check Consumer Expiry), a transformative deep learning framework designed to automate inventory tracking and shelf-life monitoring. For packaged goods, the system implements a two-stage pipeline utilizing a Convolutional Recurrent Neural Network (CRNN) integrated with Vision Transformers (ViT) to extract expiry dates with high precision. For unpacked fresh produce, the framework employs a CRNN-based ensemble model and a Dynamic Mathematical Prediction Engine to estimate shelf life by fusing visual freshness scores with real-time environmental data, such as ambient temperature and humidity. By synchronizing shop-owner and customer applications, the platform provides proactive 24-hour expiry alerts and automated "Digital Pantry" updates. Experimental results demonstrate that PACE achieves a 92.4% word accuracy in transcribing industrial fonts—an 824% improvement over traditional benchmarks—and reduces overall food waste by up to 40%. This scalable solution fosters a sustainable supply chain while ensuring a higher standard of global consumer safety.

Fields of Study & Subjects

Handwritten Text Recognition Techniques Advanced Neural Network Applications Food Supply Chain Traceability

Publication details

Journal / source
International Journal For Multidisciplinary Research
Volume
8
Issue
2
Publication date
27 March 2026
Language
EN