Hello, I tried antsMultivariateTemplateConstruction2.sh as well as buildtemplateparallel.sh butI don't understand why the following error is occurred, only the average template is constructed. I didn't use any initial template.
Also,do you think it's logical to register subjects again with each other to match the blood vessels even after they are registered to template space?
Actually, my data is 4D, each subject containing 19 volumes. Would you mind suggetsing me a good article or paper to construct a template from 4D data using 12/13 subjects' brain, since I am not getting the point very well without details. And I can use antsMultivariateTemplateConstruction script for this purpose, right?
hello, I am confused if we can match blood vessels by directly registering a subject to a template rather than registering with another subject what I was doing(already registered to template but the vessels aren't matching that's why registered to another subject after skull removal) Would you please check the below template image, I didn't make it myselt, it looks ok but when I remove skull and projects into 2D , it doesn't show the vessels that way, is it because of the overlap of 12 subjects...
Hi, thanks a lot for all of yours kind help. The registration I got according to your suggestion is working for me better than before, still I'm trying to improve the result of small vessels. But I'm really grateful to you all. I'll write again if I face any further problem.
Hi, I tried with multiple landmarks also,but it doesn't change anything. Do you think it's not possible to match them? Landmark selection takes a lot of time, in this case the process isn't automatic. Is there any script in ANTs to choose the landmark automatically and save the segmented image? Because I have around 80 subjects' image and it's not an optimal way to choose landmark for each separately.
I'm thinking if feature-based registration would work for small vessel matching. In that case, should I use github script of feature matching by stnava?
Also I wanted to mention that the registration is better than before but it's not perfect. For my work, it should be better. Would you mind suggesting some further steps that I can perform to get better result