Metadata-Version: 2.1
Name: roi-tanh-tensorflow
Version: 0.0.2
Summary: apply roi tanh warping in tensorflow for face parsing
Author-email: Edouard Yvinec <ey@datakalab.com>
Project-URL: documentation, https://openaccess.thecvf.com/content_CVPR_2019/papers/Lin_Face_Parsing_With_RoI_Tanh-Warping_CVPR_2019_paper.pdf
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.9
Description-Content-Type: text/markdown

# roi_tanh_warping_tensorflow


This is a tensorflow re-impelementation of the [Face Parsing With RoI Tanh-Warping](https://openaccess.thecvf.com/content_CVPR_2019/papers/Lin_Face_Parsing_With_RoI_Tanh-Warping_CVPR_2019_paper.pdf) paper published at CVPR 2019. This is intended to be used for face-parsing training in tensorflow.

## Install

you can use `pip install roi-tanh-tensorflow`
## How to use

The main function to call from `src.tanh_warper` is `apply_roi_tanh_warping`. This function works as follows
```python
from roi_tanh.tanh_warper import apply_roi_tanh_warping

apply_roi_tanh_warping(image: np.ndarray, label: np.ndarray, size:Tuple[int])
```

The expected input are a single image and its corresponding label.

## run a unit test
```
python -m src.roi_tanh.tanh_warper --unit_test
```
