Sequential Monte Carlo methods for inverse estimation in imaging scienceMS74

Sequential Monte Carlo (SMC) methods provide an attractive way to estimate unknown quantities from noisy observations, as they are flexible, easy-to-implement and have reduced computational burden compared to Markov Chain Mote Carlo methods. Last decade has witnessed an explosion of scientific articles on SMC methods and their applications to inverse problems and image reconstruction, thanks to the ever-increasing computing power. The aim of this minisymposium is to bring together experts working in this field and introduce the latest algorithmic developments and applications of SMC methods pertaining to multiple target tracking, brain connectivity estimation and multimodal sensor analysis, among others.

Fri 08 June at 14:00 Matemates (Matemates, floor 0)
Expectation--maximization algorithm with a Rao-Blackwellized particle smoother for joint estimation of neural sources and connectivity from MEG data
Narayan Puthanmadam Subramaniyam (Aalto University)
Tracking of cyclists in the velodrome using IMU and timing sensors
Jiaming Liang (University of Cambridge)
Rao-Blackwellized particle filtering in multiple target tracking
Simo Särkkä (Aalto University, Department of Electrical Engineering and Automation)
Bayesan sequential Monte Carlo approaches to simulated EEG-fMRI and EEG-fNIRS data
Filippo Zappasodi ("G.d'Annunzio" University, Chieti)
Narayan Puthanmadam Subramaniyam (Aalto University)
Sara Sommariva (Aalto University)
bayesian filtering, bayesian methods, particle filters, sequential monte carlo, statistical inverse estimation methods