Open ShaodeYuOutlook opened 5 years ago
In particular, using the same data as the input, my result is totally different from the result of http://quicknat.ai-med.de/ . It frustrates me so much.
Hi,
There is a quick pre-processor using FreeSurefer before segmentation.
You need to execute this command from terminal: mri_convert (input_vol_name) (out_vol_name) --conform
This resamples and also performs some global contrast adjustments. The output is used as input to QuickNAT. In the web-service we have this pre-processor integrated.
If you are using Windows, I suggest maybe use a virtual machine with linux to do this step. Or use a Mac. Unfortunately there is no FreeSurfer for windows.
Hope this helps.
Are there any toolboxes to replace Freesurfer? I am working under Windows 10 OS.
The QuickNAT wrapped by MatConvNet works. I read in a volume image using Matlab and send it as the input of SegmentVol. However, the segmentation result is weird. Therefore, is it possible to make use of some other surrogate Freesurfer for pre-data processing, such as transforming data into mgz format, and to realize iso-tropic resolution.
Hope it is clear. Thanks.