Scalable magnetic resonance fingerprinting: Incremental inference of high dimensional elliptical mixtures from large data volumes

Magnetic Resonance Fingerprinting (MRF) is an emerging technology with the potential to revolutionize radiology and medical diagnostics. In comparison to traditional magnetic resonance imaging (MRI), MRF enables the rapid, simultaneous, non-invasive acquisition and reconstruction of multiple tissue parameters, paving the way for novel diagnostic techniques. In the original matching approach, reconstruction is based on the search for the best matches between in vivo acquired signals and a dictionary of high-dimensional simulated signals (fingerprints) with known tissue properties. A critical and limiting challenge is that the size of the simulated dictionary increases exponentially with the number of parameters, leading to an extremely costly subsequent matching. In this work, we propose to address this scalability issue by considering probabilistic mixtures of high-dimensional elliptical distributions, to learn more efficient dictionary representations. Mixture components are modelled as flexible ellipitic shapes in low dimensional subspaces. They are exploited to cluster similar signals and reduce their dimension locally cluster-wise to limit information loss. To estimate such a mixture model, we provide a new incremental algorithm capable of handling large numbers of signals, allowing us to go far beyond the hardware limitations encountered by standard implementations. We demonstrate, on simulated and real data, that our method effectively manages large volumes of MRF data with maintained accuracy. It offers a more efficient solution for accurate tissue characterization and significantly reduces the computational burden, making the clinical application of MRF more practical and accessible.

Domaines

Applications [stat.AP]

Fichier principal

2412.10173v1.pdf (5.59 Mo)

BIC.png (114.98 Ko)

HDDC.png (104.94 Ko)

bic_hdgmm.jpg (18.51 Ko)

bic_hdstm.png (25.21 Ko)

bic_scores_vasc.png (86.54 Ko)

comparison.png (2.1 Mo)

convol.png (153.76 Ko)

dim_hdgmm.png (112.63 Ko)

dim_hdstm.png (125.57 Ko)

err_comp.png (69.83 Ko)

err_d10_k15.png (478.77 Ko)

err_d10_k30.png (777.84 Ko)

err_d10_k30_gllim.png (179.55 Ko)

full_fig.png (1.13 Mo)

graphical_HDMMED.png (191 Ko)

mse_scores_vasc.png (92.04 Ko)

pipeline.png (1.33 Mo)

rec_d10_k15.png (639.39 Ko)

rec_d10_k30_gllim.png (442.75 Ko)

smoothing_gllim.png (240.22 Ko)

ssim.png (84.71 Ko)

sub_0_full_matching.png (1.21 Mo)

sub_0_hdstm_25.png (1.34 Mo)

Origine	Fichiers produits par l'(les) auteur(s)
Licence	Paternité

Origine	Fichiers produits par l'(les) auteur(s)

Geoffroy Oudoumanessah : Connectez-vous pour contacter le contributeur

https://hal.science/hal-04836864

Soumis le : vendredi 13 décembre 2024-15:28:10

Dernière modification le : jeudi 19 décembre 2024-03:12:33

Dates et versions

hal-04836864 , version 1 (13-12-2024)

Licence

Paternité

Identifiants

HAL Id : hal-04836864 , version 1
ARXIV : 2412.10173

Citer

Geoffroy Oudoumanessah, Thomas Coudert, Carole Lartizien, Michel Dojat, Thomas Christen, et al.. Scalable magnetic resonance fingerprinting: Incremental inference of high dimensional elliptical mixtures from large data volumes. 2024. ⟨hal-04836864⟩

Exporter

BibTeX XML-TEI Dublin Core DC Terms EndNote DataCite

Collections

UNIV-ST-ETIENNE UGA CNRS INRIA UNIV-LYON1 INSA-LYON CREATIS INSMI LJK INRIA2 GENCI INSA-GROUPE UDL ANR GRENOBLEINSTITUTNEUROSCIENCES

0 Consultations

0 Téléchargements