Feasibility of near-infrared markers for guiding surgical robots

Abstract

Automating surgery using robots requires robust visual tracking. The surgical environment often has poor light conditions where several organs have similar visual appearances. In addition, the field of view might be occluded by blood or tissue. In this paper, the feasibility of near-infrared (NIR) fluorescent marking and imaging for vision-based robot control is studied. The NIR region of the spectrum has several useful properties including deep tissue penetration. We study the optical properties of a clinically-approved NIR fluorescent dye, indocyanine green (ICG), with different concentrations and quantify image positioning error of ICG marker when obstructed by tissue.