Particle Filter Based Active Localization of Target and Needle in Robotic Image-Guided Intervention Systems.

This paper presents a probabilistic method for active localization of needle and targets in robotic image guided interventions. Specifically, an active localization scenario where the system directly controls the imaging system to actively localize the needle and target locations using intra-operative medical imaging (e.g., computerized tomography and ultrasound imaging) is explored. In the proposed method, the active localization problem is posed as an information maximization problem, where the beliefs for the needle and target states are represented and estimated using particle filters. The proposed method is also validated using a simulation study.