Surface Reconstruction for a DIET breast cancer screening system

Abstract This paper outlines a method for accurately tracking surface motion on an actuated breast as part of a Digital Image-based Elasto-Tomography (DIET) breast cancer screening project. The tracking method is validated in both simulation and experiment on a silicon breast phantom which has similar elasticity characteristics to human breast tissue. Given mean reprojection error ranges between 0.3 and 0.5 pixels for 2M pixel cameras the end result is sub-millimetre surface tracking of a silicon phantom actuated at 100 Hz. This highly accurate, fast and low cost method of tracking is thus well suited to a DIET system.