A Deep Learning-based Accurate Drug Detection, Identification and Confirmation Mechanism for Medication Dispensing Package of Inpatients

Abstract

In this study, a deep learning based accurate drug detection, identification and confirmation mechanism (DLADDICM) for medication dispensing package is proposed for inpatients. In this proposed DLADDICM, a medication dispensing package with a printed QR code is photo taken and drugs in the image are detected and identified using a deep learning object detection algorithm, namely You Only Look Once (YOLO). The QR code information is deciphered and compared with the detected drugs to confirm the correctness of the medication dispensing. If there are mismatch situation(s), the computer with the proposed DLADDICM will generate different warning sound in responding to different incorrect situations. A data set with 30 drugs form the National Library of Medicine of NIH, USA is used for testing the DLADDICM using the object tracking and detection deep learning algorithm YOLOv3. Experimental results shown that the DLADDICM can detect and identify the incorrect drugs and generate the appropriate warning sound for the incorrect drug in pharma package for further human inspection. The experimental results also exhibits that by using the AI-enabled mechanism an accurate, safer, healthier with precision medication environment for the medical industries could also be achieved.