Integrated design of an intelligent surgical gauze counting solution with commercialization potential

Abstract

Gossypiboma is a common but avoidable medical complication mainly caused by human error during surgeries. Such cases continue to occur even in today’s modern hospitals, causing unnecessary problems for both the patient and the hospital. Currently, the main way of preventing this involves medical staff physically counting gauzes during and at the end of the surgery, and even X-Ray of the patient after the surgery is employed as a secondary safety net. However, the reliability of the first solution, as well as the cost of the second solution highlight a gap in this area.

As such, project introduces an innovative AI-powered gauze counting system, aiming to minimize the risk of gossypiboma. Leveraging YOLOv5 and tailored hardware, our system enhances surgical safety by reducing manual labour and eliminating reliance on costly secondary measures like X-rays. Key contributions include the development of a traffic-light anomaly detection system, live operating theatre tests, and fine-tuning algorithms for robustness. These innovations mark a significant leap toward commercial and clinical viability.