Julien Barbot1, Steven Shea2, Klaus Kirchberg1, Kamal Vij3, Sunil Patil2
1Center for Applied Medical Imaging, Siemens corporation, Corporate Research, Princeton, NJ, United States; 2Center for Applied Medical Imaging, Siemens corporation, Corporate Research, Baltimore, MD, United States; 3MRI Interventions, Irvine, CA, United States
In order to accurately determine and estimate the micro-coil position mounted on an interventional device, a novel multi-scale tracking method is proposed by considering the projection signal of interest as a singularity smoothed by a diffusion process. The parameter estimation is then done by performing edge detection at different scales and characterizes them to find suitable patterns for a given projection signal. The resulting patterns from different scales are combined to obtain the likelihood estimation at every position. The benefits over previously proposed methods are increased robustness, low sensitivity to noise and the ability to estimate the likelihood information at every position over the field of view.