Comments (3)
pinging @mu40
from voxelmorph.
This is certainly a challenging case for the affine SynthMorph model, which (1) hasn’t seen the FOV extend so much below the neck at training and (2) has been trained primarily on full brains. Your best bet would be retraining with CT-derived label maps that extend as far down as the CT scan you provided.
Barring that, you can try downloading and passing the experimental affine weights from w3id.org/synthmorph to the latest SynthMorph using the -w
flag. While this model has been trained on randomly cropped brains, I suspect your example is still at (or beyond) the limit of what that model can do.
Things that may help the experimental model include cropping the CT scan around the head to a region of similar size as the MRI (updating the voxel-to-world matrix appropriately) and clipping the CT intensities to something like [0, 80] in Hounsfield units.
from voxelmorph.
I see. Thank you very much!
from voxelmorph.
Related Issues (20)
- About the generation process of flow field/displacement field HOT 2
- affine transform from voxelmorph not compatible with ITK affine transform HOT 1
- SynthMorph shapes demo - change of loss functions HOT 3
- Need a tutorial for my own image HOT 6
- Train my own Voxelmorph model to align 3D images to mni152
- pytorch pre-train file HOT 8
- Any training tutorials? HOT 4
- Register cropped image to non-cropped image HOT 1
- Resulting VoxelMorph Atlas is Blurry and Lacks Detail HOT 6
- VoxelMorph Atlas is just a slightly blurred version of the fixed image HOT 23
- Mouse Brain Registration (rsfmri - MRI - T2w)
- Running configuration for other brain datasets with non-standard sizes
- The application of synthmorph to the problem of large deformation of lungs in ct-cbct HOT 1
- huge loss HOT 2
- How to specify input/output pairs in /images/list.txt for model training HOT 2
- How to apply the affine matrix from SynthMorph HOT 5
- How to deform annotated landmarks with the predict dense deformation field in the pytorch version
- Torch train.py problem HOT 2
- L smooth Loss not decreasing
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from voxelmorph.