Integration of a Smart Medicine Container with Medicine Sorting using YOLOv4 and OCR

Senior citizens are taking medicines on a regular basis; therefore, .it is highly probable for them to forget to drink their medications at the right times or to drink the right classification or dosage of medications. To solve this problem, several researchers made smart pill box systems that would remind their users when they should take their medications and what medication to take. As such, each of them only focused on some or most of the features stated: (a) automatic dispensing, (b) user notification, and (c) checking if the dispenser still has medicine stored while utilizing their own methodologies. Thus, the researchers of this study aimed to develop a Smart Medicine Container with Medicine Sorting using the algorithms YOLOv4 and OCR. After testing, it was found that using the YOLOv4 algorithm only is the most accurate in classifying medications for sorting (95.2381% accuracy) followed by YOLOv4 with Tesseract OCR (19.0476% accuracy) and Tesseract OCR only (14.2857% accuracy). On the other hand, the sorting capabilities of the system together with its medication provision features were 100% accurate with proper calibration of the system’s stepper motors.