Tracking system for magnetic foreign bodies localization using a portable device

Abstract

This paper presents a tracking system to find the position and yaw rotation of a handheld device based on the imaging of an adhesive marker attached to a person's body. Our goal is to provide accurate data that can be combined with data from a giant magnetoresistance sensor to be able to locate a foreign body inside a human body for surgical removal. We tested the system with a series of computer simulations and analyzed the results obtained. Our best setup provided a Root Mean Square Error of less than 1.6 mm for the x coordinate, less than 1 mm for the y coordinate, less than 1.2 cm for the z coordinate and less than 1.1° for the yaw rotation value in a test with wide variations of all values. The results obtained are promising and open the way for the combination of data with the portable magnetic sensor.

Keywords

• Biomagnetism
• Camera pose
• Magnetic foreign body
• GMR
• Data fusion
• Convolutional neural networks