BUG: The segment of two adjacent layers is identical

[goodsave]

hello~
After nnUNetV2 (3d_fullres) uses the same data for training, there will always be two adjacent slice of segmentation that are identical, but in fact they should be different. I guess it may be caused by segmentation after resampling CT. How can we avoid this problem?

Best wishes！

[faridyagubbayli]

I encountered a similar problem and solved by specifying lower spacing when processing the data. Example command I used:

nnUNetv2_plan_and_preprocess -d 001 <some other params> -overwrite_target_spacing 3.0 0.9 0.9

[goodsave]

Thank you very much, I will try this parameter.