Review of Semantic Analysis of Food Labels for Expiry Date Recognition Using Deep Learning and NLP
TL;DR
An automated system to detect expiration dates using deep learning techniques is proposed and contributes to improving food security and reducing waste and supports intelligent automation in food inventory management with the help of semantic analysis on the basis of intensive learning of the product label.
Abstract
Ensuring freshness and safety of packed food products is a growing concern, especially with an increase in consumption of processed goods. Expiry dates on the product label provide an important indicator for safe use. However, discrepancies in the quality of label formats, source styles, languages and prints often monitor manual and prone to errors. This can lead to involuntary consumption of expired foods and can increase food waste. The paper proposes an automated system to detect expiration dates using deep learning techniques. The system can identify the text related to the termination of the product and separate it from other types of dates, such as manufacturing and packaging dates, computer visions and a combination of natural language processing. It is designed to handle real-world challenges, such as various date formats and packaging materials, multilingual materials, low quality images, and several concurrent dates on single labels. The review contributes to improving food security and reducing waste and supports intelligent automation in food inventory management with the help of semantic analysis on the basis of intensive learning of the product label.