You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The goal for the new features is to include the building blocks of Diffusion Model based MRI reconstruction and test-time hyper-parameter tuning, as implemented in this research repo. The new features would be added to MONAI-core. Based on the discussion with @wyli, those new features would be:
the SURE loss function (line 6 in the SMRD algorithm in the link above)
the generation step using pre-trained score function (line 3 in the algorithm)
the conjugate gradient step to enforce measurement consistency (line 4 in the algorithm)
The plan is to first implement the SURE loss and then the conjugate gradient step. For the generation step, we may already have good implementations in this Generative Models repo in MONAI. I will look into the inference module in the this repo, and start from there if new implementation is needed.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Hi everyone,
The goal for the new features is to include the building blocks of Diffusion Model based MRI reconstruction and test-time hyper-parameter tuning, as implemented in this research repo. The new features would be added to MONAI-core. Based on the discussion with @wyli, those new features would be:
The plan is to first implement the SURE loss and then the conjugate gradient step. For the generation step, we may already have good implementations in this Generative Models repo in MONAI. I will look into the inference module in the this repo, and start from there if new implementation is needed.
Please feel free to comment.
Beta Was this translation helpful? Give feedback.
All reactions