Smart Surgical Scan: A Device for Detecting Retained Foreign Objects in Operative Fields

Overview

The purpose of this project is to design a non-invasive, multi-sensor scanning device that is capable of detecting retained foreign objects embedded in the epithelial layer of the skin during surgical procedures. Unintentionally retained surgical items (RSIs) pose serious post-operative risks, such as infection, pain, and even death. Existing detection methods, such as manual counting are prone to human-error and time-consuming and costly. This project aims to provide a faster and more reliable alternative for improving surgical outcomes. The device consists of three complementary sensors: a metal detector, ultrasound sensor, and a bioimpedance sensor. Each sensor captures unique data related to material type, tissue composition, and object depth. Test objects of varying material were embedded at different depths and sensor outputs were recorded and classified using a machine learning model. Performance was evaluated by comparing detection accuracy across individual sensors and the combined system. The bioimpedance sensor achieved 89% accuracy, the ultrasound sensor produced 81% accuracy at 1mm depth and 58% accuracy at 3mm depth, and the metal detector sensor achieved 97% detection accuracy. Overall, detection became more difficult as object depth increased and material differences decreased. The impedance sensor gave the most reliable results because conductive materials create strong electrical contrast. Ultrasound became less effective at greater depths due to sound scattering. Human measurements showed more variability than gelatin because skin structure affects electrical behavior.