Comments (4)
thanks for your reply, and yes. If you open source the training code, I am willing to implement the pytorch counterpart.
from livermask.
Thanks!
from livermask.
No, but it should not be that relevant as this is just an inference tool. However, it is simple to convert the model to ONNX, which can be deployed from PyTorch. The actual inference part is only a small part of the whole pipeline.
But I would not spent time on that, unless you are interested in integrating the liver segmentation part directly into some other framework based on pytorch. Is That what you are trying to do?
from livermask.
The training of the model used in this project was extremely simple (basic 2D U-Net), and I did not think about making it sharable with others. I therefore no longer have access to the code. However, if you wish to train your own models, and you want to use pytorch, I would strongly recommend using MONAI.
MONAI has all the components you need for training models and deploying them, specifically for medical data. The framework also contains a lot of different methods relevant for 3D medical data (preprocessing of CT/MRI, patching of data, one-hot encoding, loss functions, data augmentations... You name it!). If you wish to train your own model using MONAI, I would also suggest using a 3D architecture, such as the SWIN UNETR, which is easily accessible in MONAI (information regarding the architecture can be found here).
Lastly, if you just want to train a model, without going through all the hassle of implementing all the training code yourself and all that. I would recommend MONAI-Label. The software enables you to train models (and annotate, but which is not relevant in your use case) fully through a GUI, without the need to code. It also has support for the UNETR architecture. It is also extremely easy to setup. Handles everything for you from NIFTI to preprocessing to training!
A demo video of MONAI-Label can be found on YouTube. If you have any questions regarding the software, feel free to ask me. Otherwise, just make an issue in the MONAI-Label repository. They are really fast to reply and make a great effort to assist you in your endevours.
If you have any other questions regarding architecture design, which training setup to use, and whatnot. Feel free to ask me. If you are considering doing a publication, I am also open to collaborate.
Lastly, if you end up with a model that greatly outperforms my very simple livermask tool (should not be that hard tbh). It would be great if you would keep in touch, as I would be very interested in improving the parenchyma segmentation (and vessels) in my tool. If you'd like, you could make a PR for integrating your model (with the whole MONAI framework stuff), or we could do it together. Up to you :]
I hope you found this helpful. Best of luck! :]
from livermask.
Related Issues (11)
- Segmentation is not performed - possible typo in Livermask.py? HOT 4
- Fine tuning on new images - input shape? HOT 1
- ImportError: numpy.core.multiarray failed to import HOT 3
- Model downloads fail HOT 1
- MissingSchema issue HOT 9
- Verbose HOT 2
- Dependencies collision
- GitHub Releases HTTP download issue
- Error when installing livermask in windows 10 HOT 3
- Model no longer accessible HOT 3
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 livermask.