|
--- |
|
license: mit |
|
--- |
|
|
|
# BeautyBank-Inference-Dataset |
|
|
|
This dataset is used for inference in the paper: |
|
π **"BeautyBank: Encoding Facial Makeup in Latent Space"** |
|
[π arXiv:2411.11231](https://arxiv.org/abs/2411.11231) |
|
|
|
## π About the Dataset |
|
This dataset provides reference images and their corresponding **makeup codes**, enabling inference using the **BeautyBank** model. |
|
It allows users to apply various makeup styles without requiring training data. |
|
|
|
## π Dataset Contents |
|
- π **BeautyBank-Inference-Dataset.zip** β Contains **1289 reference images (PNG format)** for selecting makeup styles. |
|
- π **refined_makeup_code.npy** β Encoded makeup features for inference. |
|
|
|
## π₯ Download and Usage |
|
|
|
Since the images are stored in a ZIP file, you need to extract them before use: |
|
```bash |
|
unzip BeautyBank-Inference-Dataset.zip -d makeup_images/ |
|
``` |
|
|
|
## Citation |
|
If you use this dataset, please cite our paper: |
|
```bibtex |
|
@article{lu2024beautybank, |
|
title={BeautyBank: Encoding Facial Makeup in Latent Space}, |
|
author={Lu, Qianwen and Yang, Xingchao and Taketomi, Takafumi}, |
|
journal={arXiv preprint arXiv:2411.11231}, |
|
year={2024} |
|
} |
|
``` |