Hello.

I wonder if the output of WhyD_postproc is itself empty, i.e. the input the the quantification step are images with no/erroneous WMH detections ? Please check the 'pmap' and 'unrectify' files themselves to see if this is what is going on ?
Else, it might be the case that different images have different bias fields, which is unlikely but might cause these issues. If so, changing pre-processing hyper-parameters may help resolve the issue.
Please let me know if either of these solve the issue! Thanks for your interest.

Vamsi