calc_cartesian_plane_sampling_probability

calc_cartesian_plane_sampling_probability#

pypulseqpp.calc_cartesian_plane_sampling_probability()[source]#

Return the probability that make_cartesian_plane_sampling() acquires each view.

The arguments are those of make_cartesian_plane_sampling(). A lattice is deterministic, so its probabilities are 0 or 1. A Poisson-disc draw is replaced by its expected density, at a fraction of the cost of a draw: the density of views at minimum distances r_y and r_z, in views, is approximated as min(0.74, 0.8 / (r_y r_z)), constants fitted to draws of make_poisson_disc_mask(), with the slope of the distances set by bisection so that the expected count reaches the acceleration. The fullest partition of the density holds the lines of the fullest partition of a mean over draws to within a few per cent; a single draw departs from it by its own fluctuation.

Returns:

  • calibration (numpy.ndarray) – Probability, 0 or 1, that each view (y, z) is a calibration view.

  • imaging (numpy.ndarray) – Probability that each view is one of the other acquired views.

Examples

>>> import pypulseqpp as pp
>>> calibration, imaging = pp.calc_cartesian_plane_sampling_probability(
...     (4, 4), acceleration=(2, 2), n_acs=(2, 2)
... )
>>> int(calibration.sum()), int(imaging.sum())
(4, 3)