IOP Publishing, Inverse Problems, 2(24), p. 025023, 2008
DOI: 10.1088/0266-5611/24/2/025023
Full text: Download
A Rao–Blackwellized particle filter for the tracking of neural sources from biomagnetic data is described. A comparison with a sampling importance resampling particle filter performed in the case of both simulated and real data shows that the use of Rao–Blackwellization is highly recommended since it produces more accurate reconstructions within a lower computational effort.