A novel stock counting system for detecting lot numbers using Tesseract OCR

Abstract

Counting stock is one of the warehouse’s methods for preventing insatiable stock. Moreover, it could help the company forecast how many products they need to store and predict the replenished goods for customers. However, stock count in the medical business, which sells specialized medical equipment, needs more focus on, because it uses to treat the patient. So that lack of inventory should not happen. In a normal situation, stock count at some hospitals is quite hard for salespeople, especially hospitals in upcountry that far away. During the COVID-19 situation, many limits need to be strict. At this point, it causes a shortage of goods in many hospitals. In this paper, we represent how computer vision can help this process. When the hospital’s officer sends images of stock to our system. The system will recognize the quantity and lot number of goods that remain in the hospital. Therefore, salespeople can decrease the times to visit hospitals. The result showed that for text detection and text recognition in a specific use case. Our prototype system achieves 84.17% in accuracy.