Some error when I use post_proc.py about ribseg HOT 8 CLOSED

sitk 2.0.2 should work, the error is caused by the wrong type of input. Please double-check the input lmage_array, and make sure the data type of its source array is int, i.e., input = input.astype(int), before excuting sitk.GetImageFromArray(input).

from ribseg.

sitk 2.0.2 should work, the error is caused by the wrong type of input. Please double-check the input lmage_array, and make sure the data type of its source array is int, i.e., input = input.astype(int), before excuting sitk.GetImageFromArray(input).

HI，i have the same error，the type of its source array is int，how can i solve it？

from ribseg.

Thanks for your interest in our work.
I'm not sure why you may have this error, would you please send me one of your input files and the corresponding code? It's hard for me to debug without more info.
My email: [email protected]

from ribseg.

@1640477469 @wo-sour Fixed. The function sitk.BinaryDilate seems updated, changing the radius parameter from 3 to (3,3,3) should work, you can also try different kernel parameters for better results. Thanks for pointing out the bug.

from ribseg.

@1640477469 @wo-sour Fixed. The function sitk.BinaryDilate seems updated, changing the radius parameter from 3 to (3,3,3) should work, you can also try different kernel parameters for better results. Thanks for pointing out the bug.

Thanks again for your patience! The bug can also be fixed by using sitk 1.2.4.
However, when the rib segmentation is done, the result is not as good as yours: the rib surface is really uneven. I've tried a lot of methods to make the surface smooth, but the result is still not satisfactory. Could you give me some advice?

from ribseg.

from ribseg.

Try combination of morphological operations as the code commented. Repeated open and close operation should yield some nice results. This is a tricky part, a decent dice metric valve does not necessarily mean a visually satisfying result.

Thanks a lot! I'll try my best.

from ribseg.

I want to convert the point cloud data prediction file into a NII file. How can I modify the code? Thank you very much!

from ribseg.