This is the PyTorch implementation of Retrieve in Style: Unsupervised Facial Feature Transfer and Retrieval.

Abstract:
We present Retrieve in Style (RIS), an unsupervised framework for fine-grained facial feature transfer and retrieval on real images Recent work shows that it is possible to learn a catalog that allows local semantic transfers of facial features on generated images by capitalizing on the disentanglement property of the StyleGAN latent space. RIS improves existing art on:

1. feature disentanglement and allows for challenging transfers (\ie, hair and pose) that were not shown possible in SoTA methods.
2. eliminating the needs for per-image hyperparameter tuning, and for computing a catalog over a large batch of images.
3. enabling face retrieval using the proposed facial features (\eg, eyes), and to our best knowledge, is the first work to retrieve face images at the fine-grained level.
4. robustness and natural application to real images. Our qualitative and quantitative analyses show RIS achieves both high-fidelity feature transfers and accurate fine-grained retrievals on real images. We discuss the responsible application of RIS.

Our codebase is based off stylegan2 by rosalinity.

conda install --yes -c pytorch pytorch=1.7.1 torchvision cudatoolkit=<CUDA_VERSION>
pip install tqdm gdown scikit-learn scipy lpips dlib opencv-python

Everything to get started is in the colab notebook.

If you use this code or ideas from our paper, please cite our paper:

@InProceedings{Chong_2021_ICCV,
    author    = {Chong, Min Jin and Chu, Wen-Sheng and Kumar, Abhishek and Forsyth, David},
    title     = {Retrieve in Style: Unsupervised Facial Feature Transfer and Retrieval},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2021},
    pages     = {3887-3896}
}

This code borrows from StyleGAN2 by rosalinity, Editing in Style, StyleClip, PTI. Encoder used is borrowed directly from encoder4editing.