Deep Learning & Computer Vision based Medicine Inventory Monitoring
Case Study
Deep Learning & Computer Vision based Medicine Inventory Monitoring
Real time inventory updates and notifications, 91% mAP (Mean Average Precision) – model accuracy of medicine vials, injections classification and counting.
Executive Summary
The client is a leading US based autonomous pharmacy solutions company offers medication management solution with combination of automation, intelligence, and technology-enabled services. The client wanted to replace the existing manual medicine counting process in critical areas such as operating room, pharmacies with computer vision based automated medicine counting solution. This will help their customers to achieve automated inventory control, get real-time notifications to just in-time replenishment and improve hospital staff operational efficiency.
Client
The client is a US based leading and innovative medication management and pharmacy solutions company provides automation and business analytics solutions for patient-centric medication and supply management across the entire healthcare continuum.
Challenge
As a part of new strategic initiative, client wanted to build computer vision-based, automated medical inventory counting solution for hospital operating rooms and pharmacies. There were different medications in drawers to be monitored such as vials, bottles, injections, medicine cream tubes, medicine boxes, etc. The challenge was to capture and tune low light image input with field of view occlusion and perform accurate real-time detection, identification, counts and orientations for
multiple objects.
