-
-
Notifications
You must be signed in to change notification settings - Fork 381
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix joint fusion bugs multi modal data #1302
Fix joint fusion bugs multi modal data #1302
Conversation
For multiple modality images, the atlases were incorrectly doubled. For example, 3 atlases with T1 & T2 modalities should result in a list of lists like 0: [A1-T1, A1-T2] 1: [A2-T1, A2-T2] 2: [A3-T1, A3-T2] but instead resulted in 0: [A1-T1, A1-T2] 1: [A1-T1, A1-T2] 2: [A2-T1, A2-T2] 3: [A2-T1, A2-T2] 4: [A3-T1, A3-T2] 5: [A3-T1, A3-T2] Only the first 3 elements are processed, so about 1/2 the atlases are not used.
Scaling the target, but not the atlases results in invalid MSE minimized results when searching local regions. TARGET[i] = [ 1, 1, 1, 1] NORMALIZED_TARGET[i]=[ 0, 0, 0 ,0] ATLAS[i] = [ 1, 1, 1, 1] ATLAS[i+1]= [ 0, 1, 1, 1] MSE (NORMALIZED_TARGET[i], ATLAS[i]) = 4 MSE (NORMALIZED_TARGET[i], ATLAS[i+1]) = 3
Thanks again, @hjmjohnson . This is great. |
@hjmjohnson , I'm seeing the same thing on the failed Circle-CI build and on my local machine. Any suggestions?
|
@ntustison I'm looking right now. I think I may have compiled the wrong ant's directory with a different codebase for testing, and then pre-maturely pushed. |
@ntustison I fixed the obvious syntax errors. I don't usually build with VTK, so those errors are going to take me a few minutes to build VTK and identify. PR arriving shortly (at least for the non VTK mistakes). |
Two bugs were identified when using uncommon parameters for ants joint fusion.