Endoscope 3-D motion tracking using an aggressive particle filtering for boosting electromagnetic guidance endoscopy

The paper proposes a novel endoscope motion estimation method that bases an aggressive particle filter (APF) for enhancing electromagnetic tracking (EMT) during guided endoscopy. We explore an APF strategy to resolve two main limitations of EMT sensor measurements: (1) inaccuracy due to airway deformation and (2) instability or jitter errors because of magnetic field distortion. During such a strategy, a swarm intelligent technique - particle swarm optimization that was modified to traceably determine evolutionary parameters in this paper was embedded into a generic particle filter (GPF) to address particle impoverishment that usually occurs in GPF. Experimental results from validating our method on a dynamic bronchial phantom demonstrate our method´s effectiveness and robustness during EMT-guided endoscopy. The tracking pose error was significantly reduced from (4.3, 7.8) to (2.7 mm, 5.6°).