parker_weighting¶

odl.tomo.analytic.filtered_back_projection.parker_weighting(ray_trafo, q=0.25)[source]¶

Create parker weighting for a RayTransform.

Parker weighting is a weighting function that ensures that oversampled fan/cone beam data are weighted such that each line has unit weight. It is useful in analytic reconstruction methods such as FBP to give a more accurate result and can improve convergence rates for iterative methods.

See the article Parker weights revisited for more information.

Parameters

ray_trafoRayTransform: The ray transform for which to compute the weights.
qfloat, optional: Parameter controlling the speed of the roll-off at the edges of the weighting. 1.0 gives the classical Parker weighting, while smaller values in general lead to lower noise but stronger discretization artifacts.

Returns

parker_weightingray_trafo.range element