MONAI MetaTensor
A subclass wrapper of torch.Tensor in MONAI that carries metadata such as a spatial coordinate system (affine), original channel descriptions and voxel spacing. It is commonly used in medical image segmentation.
Timeline (last 90 days)
DebugSlicer supports PyTorch tensors, MONAI MetaTensor, NumPy arrays, and PIL images.
MONAI MetaTensor is a subclass wrapper of torch.Tensor that retains tensor computation capabilities while attaching metadata such as the spatial coordinate system (affine transformation matrix), original channel descriptions, and voxel spacing.
When debugging, looking only at values without looking at metadata makes it easy to miss "axis flips" caused by orientation matrix errors.
All Facts (3)
DebugSlicer supports PyTorch tensors, MONAI MetaTensor, NumPy arrays, and PIL images.
50%UnverifiedMONAI MetaTensor is a subclass wrapper of torch.Tensor that retains tensor computation capabilities while attaching metadata such as the spatial coordinate system (affine transformation matrix), original channel descriptions, and voxel spacing.
50%UnverifiedWhen debugging, looking only at values without looking at metadata makes it easy to miss "axis flips" caused by orientation matrix errors.
50%