Expiry Date Character Recognition on Canned Goods Using Convolutional Neural Network VGG16 Architecture

Abstract

Expiration dates have a significant role in the health and safety of consumers. It declares when a processed product like canned goods should be consumed. Once it has passed the expiration date, the product turns into food waste. This has been a major concern globally wherein one-third of the food manufactured is wasted. Using the researchers' dataset, the study focused on creating a system that uses image processing methods to recognize the characters, and the dot-matrix print expiration dates, in canned goods using the prescribed format: day, month, and year via Convolutional Neural Network Architecture. The process of image preprocessing, recognition, extraction, and classification of expiration date characters from the input image will take place in the raspberry pi microcontroller. An LCD screen was utilized to display the translated expiration dates from sample canned goods. The research used 30 sample canned goods, 27 of them were recognized correctly while the 3 others are recognized incorrectly. With this, the system was able to achieve 90% accuracy and proved that the system can recognize the expiration date on canned goods.