|
Notes:
Enclosed are some examples of simulated diffusion-weighted datasets
generated
using the framework described in:
doi:10.1016/j.neuroimage.2015.11.006
The simulations represent a typical HARDI style dataset, consisting
of 108
images with 12/32/64 volumes with b=0/700/2000 s/mm^2. The
simulations
are divided into:
-- A ground truth dataset with no artefacts and no added noise
-- Datasets with eddy currents and "small" amounts of motion at
SNR=20 and 40
-- Datasets with eddy currents and "large" amounts of motion at
SNR=20 and 40
The datasets are intended to give a feel for the sort of data that
can be
generated with the framework. The framework can also be used to
create
images with various artefacts, including ghosts, gibbs ringing,
susceptibility, signal dropout and RF spikes. Whilst the code for
the
framework is not (yet) available, I would be happy to collaborate
if you
were interested in using simulations of datasets with artefacts.
You can email me at mark.graham.13 [at] ucl.ac.uk
Changes:
|
