Brett Clark

As I understand it, the transform created by Unigradicon consists of composite: affine, DVF, and affine transforms. The outer affines project fixed image points into network coordinates (spacing: 1, offset:...

In any case, there are some tricky things when working with ITK. I'd take a look at these and see if any apply: - There's a difference between how nibabel...

Hi, thanks for creating this great fine-tuning tutorial! I'm fine-tuning on a local dataset which I'm loading from nifti. I'm just querying what format is expected for the input data...

Thanks @HastingsGreer for the quick reply. Yes, I already have a data processing pipeline which I'm using to evaluate UGI on my local dataset, and have fine-tuned a model using...

Agreed, this would be a great addition to this tool as many papers show registration improvements when segmentation labels or keypoints/landmarks are provided during training (e.g. Voxelmorph paper - https://arxiv.org/abs/1809.05231)....

Hi @sten2lu, thanks for your reply! To clarify, the issue is that I'm training on multi-class labels, for which a single sample (patient) may have missing classes (e.g. no Brain...

Hi @sten2lu, Thanks for your reply. Yes, your understanding is correct. In this instance I have whole CT images of the head and neck region, and each patient only has...

Thanks @sten2lu, It's unfortunate that you're not entertaining this as a feature. Multi-organ datasets generated through clinical practice typically have missing labels, as perhaps only a subset of structures close...