Accelerating Dynamic MRI Via Spatially Varying Causal Windows

Uygar Smbl¹, Juan Manuel Santos, John Mark Pauly¹

A causal, pixel-dependent exponential decay window is suggested to improve time series reconstruction. The study is motivated by the observation that many image pixels change slowly over time, while a few pixels experience rapid changes. The window interpretation is realized via a Kalman filter based algorithm. This fast statistical algorithm decreases the temporal blur of the sliding window reconstruction. Moreover, the algorithm handles arbitrary readout trajectories and multiple coils naturally.