https://github.com/pfcclab/deepali
About Image, point set, and surface registration in Paddle.
https://github.com/pfcclab/deepali
image-registration medical-imaging spatial-transformer-networks
Last synced: 11 months ago
JSON representation
About Image, point set, and surface registration in Paddle.
- Host: GitHub
- URL: https://github.com/pfcclab/deepali
- Owner: PFCCLab
- License: apache-2.0
- Created: 2024-11-04T09:16:25.000Z (over 1 year ago)
- Default Branch: develop
- Last Pushed: 2025-01-16T10:03:27.000Z (about 1 year ago)
- Last Synced: 2025-01-29T03:57:43.076Z (about 1 year ago)
- Topics: image-registration, medical-imaging, spatial-transformer-networks
- Language: Python
- Homepage:
- Size: 608 KB
- Stars: 0
- Watchers: 6
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
- Citation: CITATION.cff
Awesome Lists containing this project
README
# deepali(Paddle Backend)
> [!IMPORTANT]
> This branch(paddle) experimentally supports [Paddle backend](https://www.paddlepaddle.org.cn/en/install/quick?docurl=/documentation/docs/en/develop/install/pip/linux-pip_en.html)
> as almost all the core code has been completely rewritten using the Paddle API.
>
> It is recommended to install **nightly-build(develop)** Paddle before running any code in this branch.
Install:
``` shell
# paddlepaddle develop
python -m pip install --pre paddlepaddle-gpu -i https://www.paddlepaddle.org.cn/packages/nightly/cu118/
# setup config file is pyproject.toml
pip install .
# test
pytest -v ./tests
```
## example code
``` python
import paddle
from deepali.core import bspline as B
kernel = B.cubic_bspline_interpolation_weights(5)
assert isinstance(kernel, paddle.Tensor)
assert tuple(kernel.shape) == (5, 4)
assert paddle.allclose(
x=kernel, y=B.cubic_bspline_interpolation_weights(5, derivative=0)
).item()
```