PET/CT registration with DIRART

[Siti_Mustaqim]

Hi;

I just downloaded DIRART toolbox and I would like to try the deformable registration between PET and CT images thorax. When I tried to register one of the technique, error happened stated as:

 

Error happens: Too many output arguments.
1: line [7] - function [setpath]
2: line [45] - function [multigrid_7]
3: line [104] - function [start_registration_from_menu]
4: line [5920] - function [Start_Registration_Menu_Item_Callback]
5: line [96] - function [gui_mainfcn]
6: line [82] - function [reg3dgui]
7: line [0] - function [guidemfile/@(hObject,eventdata)reg3dgui('Start_Registration_Menu_Item_Callback',hObject,eventdata,guidata(hObject))]

 

Could anyone explain to me what really happens?

 

Thank you.

[Brandon Walker]

Same problem here. Did you ever find a solution to this? Whenever I try to perform a deformable registration with DIRART I get the same error message:

Error happens: Too many output arguments.

1: line [7] - function [setpath]

2: line [45] - function [multigrid_7]

3: line [104] - function [start_registration_from_menu]

4: line [5920] - function [Start_Registration_Menu_Item_Callback]

5: line [96] - function [gui_mainfcn]

6: line [82] - function [reg3dgui]

7: line [0] - function [guidemfile/@(hObject,eventdata)reg3dgui('Start_Registration_Menu_Item_Callback',hObject,eventdata,guidata(hObject))]

[Brandon Walker]

I only get this when trying to perform an original Horn and Schunck optical flow registration. Does anyone know how to get around this?

[Deshan Yang]

Do you have CERR folder path above DIRART? If yes, please make DIRART
path before CERR. CERR has many old DIRART functions included.

-Deshan

> --
> You received this message because you are subscribed to the Google Groups
> "dirart" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to dirart+un...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>
>

[Brandon Walker]

Thanks! That fixed it.

-Brandon

[Brandon Walker]

Moving the CERR folder path above DIRART allowed me to run both the HN registration algorithm and demons, but it introduced a new problem. When CERR path is above the DIRART path, I perform either an original demons registration or HN algorithm but then the jacobian analysis fails with NaN's all over. Here is the output:

==================================================

Jacobian analysis results:

==================================================

Mean = NaN

STD = NaN

Median = NaN

Value <= 0: 0 percent

However, when I put CERR path below DIRART, I am able to perform at least a demons registration and get a valid jacobian analysis. But with CERR below DIRART, HN algorithm doesn't work again. So I need CERR above DIRART in order to perform a HN, but I can't figure out what is going wrong with the jacobian analysis. I did find that the code fails in "Jacobian_Analysis_Callback.m" on line 28 when mean(jac(:)) is called. It turns out that jac(:) has NaN entries towards the end of the array, so the mean fails. I can' figure out when the NaN entries are introduced though. I get the impression something is happening with motion_field_jacobian?

Have you been able to perform an original HN registration and be able to perform a jacobian analysis? Do you know how to fix this?

Thanks,

-Brandon