Hi,
I'm encountering a strange issue using SUSAN (FSL 5.0.4) with a small FWHM. I initially ran FEAT preprocessing with smoothing set to 3mm or 5mm FWHM (3mm voxel size), and noticed that the output for 3mm FWHM appeared excessively smoothed, and more so than for 5mm FWHM. I then reproduced this behavior calling SUSAN directly with the following command:
susan prefiltered_func_data_bet 134.41685475 [SIGMA] 3 1 1 mean_func 134.41685475 outputN
where prefiltered_func_data_bet is an image that has just been motion-corrected and brain-masked, and mean_func is the temporal mean of that image. I set SIGMA to 2.123 (5mm FWHM), 2, 1.5, 1.45, 1.4, 1.3, and 1.274 (3mm FWHM). The output looks reasonable down to SIGMA=1.45, but at 1.4 and lower I start to reproduce the strange behavior: the image looks more smooth at 1.4 than at 1.45, and even moreso at 1.3. Images of this behavior can be found at https://www.dropbox.com/s/qxhigzdhdflm365/HumanSUSANTest.zip?dl=0
Any idea what's going on here? thanks,
Ben
|