Dictionary-Free MR Fingerprinting Parameter Estimation Via Inverse Regression

Fabien Boux (1, 2) , Florence Forbes (2) , Julyan Arbel (2) , Emmanuel L. Barbier (1)

1 GIN - [GIN] Grenoble Institut des Neurosciences
2 MISTIS - Modelling and Inference of Complex and Structured Stochastic Systems

MR Fingerprint requires an exhaustive search in a dictionary, which even for moderately sized problems, becomes costly and possibly intractable. In this work, we propose an alternative approach: instead of an exhaustive search for every signal, we use the dictionary to learn the functional relationship between signals and parameters. This allows the direct estimation of parameters without the need of searching through the dictionary. The comparison between a standard grid search and the proposed approach suggest that MR Fingerprinting could benefit from a regression approach to limit dictionary size and fasten computation time.

Domaines: Statistiques [math.ST]; Neurosciences [q-bio.NC]

Fabien Boux, Florence Forbes, Julyan Arbel, Emmanuel L. Barbier. Dictionary-Free MR Fingerprinting Parameter Estimation Via Inverse Regression. Joint Annual Meeting ISMRM-ESMRMB 2018, Jun 2018, Paris, France. pp.1-2. ⟨hal-01941630⟩