Curtó & Zarza
Curtó is the first GAN (Generative Adversarial Networks) augmented dataset of faces. A key step towards Artificial General Intelligence (AGI).
Neil Armstrong — 'That's one small step for a man, one giant leap for mankind'.
Links to Download:
Alternate Links to Download:
Curtó & Zarza
Department of Computer Science and Engineering. The Chinese University of Hong Kong.
For more information about the dataset, visit the website:
If you use the dataset in a publication, please cite the paper below:
Curtó, J. D., Zarza, I. C., Torre, F., King, I., Lyu, M. R. High-resolution Deep Convolutional Generative Adversarial Networks. 2017.
Please note that we do not own the copyrights to these images. Their use is RESTRICTED to non-commercial research and educational purposes.
Version 1.0, released on 17/11/2017.
Version 1.1, released on 21/11/2017, add labels with attribute information.
Version 1.2, released on 19/01/2018, add supplementary samples and labels.
Version 1.3, released on 27/01/2018, add HDCGAN Synthetic Images.
Version 2.0, released on 09/02/2018, add HDCGAN HQ Synthetic Images.
- Samples (c/samples) 14,248 cropped faces. Balanced in terms of ethnicity: African American, East-asian, South-asian and White. Mirror images included to enhance pose variation.
- Labels (c/labels) JSON files with attribute information: Gender, Age, Ethnicity, Hair Color, Hair Style, Eyes Color, Facial Hair, Glasses, Visible Forehead, Hair Covered and Smile.
- HDCGAN Synthetic Images (c/hdcgan)
4,239 faces generated by HDCGAN trained on CelebA.
- Samples (c/hdcgan/samples) Original 512x512 image size.
- Resized Samples (c/hdcgan/samples_128) 128x128 image size.
- Additional Images (c/extra)
- Samples (c/extra/samples), Labels (c/extra/labels) 3,384 cropped faces with labels, Ethnicity: White.