Comments (4)
Hi @coldgemini ,
All the examples in the gallery work fine in 3D: the tutorials are written the way they are because it is easier for me to show what is happening with 2D toy examples, but there really isn't any technical limitation as far as GeomLoss is concerned. See for instance the tutorial on color transfer, which is a simple 3D point cloud registration in the unit RGB cube.
Going further, I will implement higher-level loss functions over the summer (such as varifold kernel and OT losses for 3D meshes), and write self-contained rigid or spline registration examples. Before working on it, I just have to release KeOps v1.1 first!
Note, however, that unlike KeOps, GeomLoss will not support raw numpy input: the native support of automatic differentiation by PyTorch is a key feature that allowed me to considerably trim down the GeomLoss interface. It really should be favored over numpy for any kind of intensive application, and is on the verge of becoming a new standard for computational anatomy - my main motivation.
Best regards,
Jean
from geomloss.
from geomloss.
You're welcome! Feel free to create an other issue if you have any other question.
from geomloss.
did you integrate geomloss into your lddmm registration project?
from geomloss.
Related Issues (20)
- Dual and primal loss don't align for small blur values HOT 1
- Support for half/single floating point numbers HOT 3
- Optionally return transport plans for the Sinkhorn loss HOT 3
- Wasserstein distance for p not in {1,2}
- sinkhorn_divergence for 1D images not workin HOT 4
- Hausdorff Distance HOT 1
- Has ImagesLoss ever been finished? Or is it still a WIP? HOT 4
- CUDA_ERROR_INVALID_SOURCE error when running geomloss on some GPUs HOT 1
- Can this library be used with torch.amp? HOT 1
- generic_logsumexp with larger point clouds HOT 2
- Installing geomloss fails if torch is being installed at the same time HOT 4
- Error when using the hausdorff distance HOT 1
- Very different results for Wasserstein distance compared to Gudhi HOT 9
- Gaussian MMD what is the optimal blur
- question about the error : arange: cannot compute length HOT 3
- ValueError: Maximum allowed size exceeded when only one value
- `LazyTensor` is not defined
- Compute Sinkhorn distance for desity images
- do you have any plans for low-rank sinkhorn implementation?
- Hi @heslowen! May I ask how to obtain both the loss and the transport plan at the same time?
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 geomloss.