TY - CHAP
T1 - Acquiring and Predicting Multidimensional Diffusion (MUDI) Data
T2 - An Open Challenge
AU - Pizzolato, Marco
AU - Palombo, Marco
AU - Bonet-Carne, Elisenda
AU - Tax, Chantal M.W.
AU - Grussu, Francesco
AU - Ianus, Andrada
AU - Bogusz, Fabian
AU - Pieciak, Tomasz
AU - Ning, Lipeng
AU - Larochelle, Hugo
AU - Descoteaux, Maxime
AU - Chamberland, Maxime
AU - Blumberg, Stefano B.
AU - Mertzanidou, Thomy
AU - Alexander, Daniel C.
AU - Afzali, Maryam
AU - Aja-Fernández, Santiago
AU - Jones, Derek K.
AU - Westin, Carl Fredrik
AU - Rathi, Yogesh
AU - Baete, Steven H.
AU - Cordero-Grande, Lucilio
AU - Ladner, Thilo
AU - Slator, Paddy J.
AU - Hajnal, Joseph V.
AU - Thiran, Jean Philippe
AU - Price, Anthony N.
AU - Sepehrband, Farshid
AU - Zhang, Fan
AU - Hutter, Jana
PY - 2020
Y1 - 2020
N2 - In magnetic resonance imaging (MRI), the image contrast is the result of the subtle interaction between the physicochemical properties of the imaged living tissue and the parameters used for image acquisition. By varying parameters such as the echo time (TE) and the inversion time (TI), it is possible to collect images that capture different expressions of this sophisticated interaction. Sensitization to diffusion-summarized by the b-value-constitutes yet another explorable “dimension” to modify the image contrast, which reflects the degree of dispersion of water in various directions within the tissue microstructure. The full exploration of this multidimensional acquisition parameter space offers the promise of a more comprehensive description of the living tissue but at the expense of lengthy MRI acquisitions, often unfeasible in clinical practice. The harnessing of multidimensional information passes through the use of intelligent sampling strategies for reducing the amount of images to acquire, and the design of methods for exploiting the redundancy in such information. This chapter reports the results of the MUDI challenge, comparing different strategies for predicting the acquired densely sampled multidimensional data from sub-sampled versions of it.
AB - In magnetic resonance imaging (MRI), the image contrast is the result of the subtle interaction between the physicochemical properties of the imaged living tissue and the parameters used for image acquisition. By varying parameters such as the echo time (TE) and the inversion time (TI), it is possible to collect images that capture different expressions of this sophisticated interaction. Sensitization to diffusion-summarized by the b-value-constitutes yet another explorable “dimension” to modify the image contrast, which reflects the degree of dispersion of water in various directions within the tissue microstructure. The full exploration of this multidimensional acquisition parameter space offers the promise of a more comprehensive description of the living tissue but at the expense of lengthy MRI acquisitions, often unfeasible in clinical practice. The harnessing of multidimensional information passes through the use of intelligent sampling strategies for reducing the amount of images to acquire, and the design of methods for exploiting the redundancy in such information. This chapter reports the results of the MUDI challenge, comparing different strategies for predicting the acquired densely sampled multidimensional data from sub-sampled versions of it.
KW - Diffusion
KW - MUDI
KW - Quantitative imaging
KW - Relaxation
UR - http://www.scopus.com/inward/record.url?scp=85095863984&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-52893-5_17
DO - 10.1007/978-3-030-52893-5_17
M3 - Chapter
AN - SCOPUS:85095863984
T3 - Mathematics and Visualization
SP - 195
EP - 208
BT - Mathematics and Visualization
PB - Springer Science and Business Media Deutschland GmbH
ER -