Skip to content

skylab-tech/ffhqr-dataset

main
Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?
Code

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time

Skylab Research

Flickr-Faces-HQ-Retouching (FFHQR) Dataset

License CC Format PNG Resolution 1024×1024 Images 70000 Teaser image

Flickr-Faces-HQ-Retouching (FFHQR) is a high-quality image dataset of retouched human faces. The dataset is released as part of the following paper:

AutoRetouch: Automatic Professional Face Retouching
Alireza Shafaei (skylabtech), James J. Little (UBC), Mark Schmidt (UBC)
Winter Conference on Applications of Computer Vision (WACV), 2021
WACV 21 PagePDFSuppVideo

The original FFHQ dataset consists of 70,000 1 MP face-aligned images that are collected from Flickr. We professionally retouched FFHQ to create FFHQR. FFHQR is the first large-scale publicly available retouching dataset. We chose FFHQ as the basis of our new dataset because of the variety of ages, ethnicity, lighting conditions, and the large number of images that could benefit from face retouching.

For press, business, and other inquiries, please contact alireza@skylabtech.ai

Licenses

The retouching dataset is made available under Creative Commons BY-NC-SA 4.0 license by Skylab Technologies Incorporated. You can use, redistribute, and adapt it for non-commercial purposes, as long as you (a) give appropriate credit by citing our paper, (b) indicate any changes that you've made, and (c) distribute any derivative works under the same license.

@InProceedings{Shafaei_2021_WACV,
    author    = {Shafaei, Alireza and Little, James J. and Schmidt, Mark},
    title     = {AutoRetouch: Automatic Professional Face Retouching},
    booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)},
    month     = {January},
    year      = {2021},
    pages     = {990-998}
}

Overview

To use FFHQR, you first need to download the original 1024x1024 FFHQ images.

The retouched images (FFHQR) can be downloaded below.

Path Size Files Format Description
ffhqr-dataset
├  images1024x1024 - part 1 13 GB 10,000 PNG (tar package) 00000-09000
├  images1024x1024 - part 2 13 GB 10,000 PNG (tar package) 10000-19000
├  images1024x1024 - part 3 13 GB 10,000 PNG (tar package) 20000-29000
├  images1024x1024 - part 4 13 GB 10,000 PNG (tar package) 30000-39000
├  images1024x1024 - part 5 13 GB 10,000 PNG (tar package) 40000-49000
├  images1024x1024 - part 6 13 GB 10,000 PNG (tar package) 50000-59000
├  images1024x1024 - part 7 13 GB 10,000 PNG (tar package) 60000-69000
└  thumbnails128x128 2.3 GB 70,000 PNG (tar package) Thumbnails at 128×128
  • Data Split (by folders):
    • Train: 00000 to 55000
    • Validation: 56000 to 62000
    • Test: 63000 to 69000

Acknowledgements

We would like to thank the PhotoRetouchOnline.com team and the Artona Group Inc. staff for their valuable feedback and support. We thank Tero Karras (NVIDIA), Samuli Laine (NVIDIA), Timo Aila (NVIDIA) for the original FFHQ dataset.

Privacy

If you wish to remove a photo that you own from FFHQ and FFHQR, please follow the FFHQ instructions. After your image is successfully deleted from FFHQ, please contact alireza@skylabtech.ai to remove the corresponding image from FFHQR.

About

FFHQR -- the first large-scale retouching dataset for computer vision research.

Topics

Resources

License

Stars

Watchers

Forks