compute_histograms is malfunctioning

jamesavery / maxibone

SRμCT Bone tomography analysis

0 stars 0 forks source link

compute_histograms is malfunctioning #23

Closed jamesavery closed 5 months ago

jamesavery commented 2 years ago

At some point, an update killed the field histogramming computation in compute_histograms.py/histograms.cc. The field histogram currently bears no relation to the true ones (EDT looks like the same distributions independent of EDT value).

Suspect: Indexing block with full-dimensions vs block dimensions. Also candidate for rewrite (simplify + GPUize)

jamesavery commented 2 years ago

Main problem fixed, but there is now a weird artefact in the EDT histogram:

fb-edt-bone_region0

carljohnsen commented 5 months ago

Fixed during #25 - the edt field now looks like this: