Hello,
This is a feature request, not a bug.
In the simple region growing module, there is the multiplier parameter which is to my understanding, unless I'm mistaken :
mean_intensity +/- (multiplier * std deviation).
Does an asymmetric multipler make sense with regards to your algorithms ?
mean_intensity + (upper_bound_multiplier std deviation)
mean_intensity - (lower_bound_multiplier std deviation)
The idea is to fine tune voxel selection in situations where intensities are idely distributed.
My primary concern is dealing with CT angiograms. Simple region growing is an invaluable tool for model creation. Sometimes I need to include an arterial branch, like an inferior mesenteric artery, where intensities are quite lower than in the aorta, even if the exam is well performed. If I include a fiducial point directly on that branch, then there are too many non arterial structures where intensities are low that are selected. The assumption is that if the lower bound multiplier can be reduced, the resulting label map would be more precise.
Thank you for considering. | 
