Recommendation of Product Value by Extracting Expiry Date using Deep Neural Network

Abstract

The quality of food products plays a vital role in maintaining the standards of human health. A large quantity of expired food products is thrown away as it is not safe to consume. Proper labeling of expiry dates ensures both the safety and the quality of food. In most of the packaged food products, the manufacturing date of the product is available along with the shelf-life period of the product. In this paper, we propose an artificial intelligence-based approach to extract and perceive the expiry date in packaged food for estimating its remaining shelf life by fetching the relevant fields from the inventory database. Based on the remaining shelf-life discount rates are calculated for the food products to promote the sales with attractive discount rates for the consumer and reduced food wastage for the seller. The proposed approach performs two-level processing for date recognition from the product's scanned images. Initially, it performs object detection using the Single Shot Detector MobileNet (SSD MobileNet) model to extract the boundaries of the date from the scanned image. Following the extraction of date, the text in it is recognized using the Attention OCR model. On extensive evaluation, the proposed method of date extraction shows better performance than other models with nearly 3% improved accuracy.