Overlay two labels into a dual channel label

[wxr521314]

I have data.nii.gzl. Label.nii.gz and label_2.nii.gz, I would like to add label.nii.gz, Label_2.nii.gz is superimposed into a two channel data. Can LoadNiftid be automatically completed? What should I do?

[wxr521314]

I have a dual channel 128 * 128 * 128 data, and I use EnsureChannelFirst(), torch.Size ([1, 2, 128, 128, 128, 1]), why is there an extra dimension? It shouldn't be torch.Size ([1, 2, 128, 128, 128,]) correct?

[KumoLiu]

I have data.nii.gzl. Label.nii.gz and label_2.nii.gz, I would like to add label.nii.gz, Label_2.nii.gz is superimposed into a two channel data. Can LoadNiftid be automatically completed? What should I do?

You can use ConcatItemsd to help you concate.

MONAI/monai/transforms/utility/dictionary.py

Line 927 in e5afa43

class ConcatItemsd(MapTransform):

I have a dual channel 128 * 128 * 128 data, and I use EnsureChannelFirst(), torch.Size ([1, 2, 128, 128, 128, 1]), why is there an extra dimension? It shouldn't be torch.Size ([1, 2, 128, 128, 128,]) correct?

EnsureChannelFirst is used to ensure the channel dim first, if you already have 2 as your channel dim, no need to add this transform.

Thanks.
(Move to discussion as more like a usage issue.)

[wxr521314]

If I don't use EnsureChannelFirst(),torch.Size is ([1, 128, 128, 128, 1, 2]). Is that right? Using nibabel to read a multi-channel nii.gz, does the result automatically add one more dimension?