First Impressions V2 (CVPR'17) - predicted_attributes_WACV2011

Data description

The ChaLearn First Impression (FI) Dataset: http://chalearnlap.cvc.uab.es/dataset/24/description/

IMPORTANT NOTES:

- Predicted values (Attractiveness score and perceived age) are provided for a subset of the FI dataset (Caucasian), based on Ethnicity annotations already provided with the data.
- Predicted values were used as soft labels, and used as proof of concept in \cite{Junior_2021_WACV}.
- Attractiveness range from 1 to 5.

If you use these data, plese refer the folowing paper:

@InProceedings{Junior_2021_WACV,
    author    = {Junior, Julio C. S. Jacques and Lapedriza, Agata and Palmero, Cristina and Baro, Xavier and Escalera, Sergio},
    title     = {Person Perception Biases Exposed: Revisiting the First Impressions Dataset},
    booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops},
    month     = {January},
    year      = {2021},
    pages     = {13-21}
}

======================================

- Gender labels (1=male; 2=female) were manually annotated as "perceived gender".

If you use the gender labels, please refer the folowing paper:

@ARTICLE{8999746,
  author={H. J. {Escalante} and H. {Kaya} and A. A. {Salah} and S. {Escalera} and Y. {Güç;lütürk} and U. {Güçlü} and X. {Baró} and I. {Guyon} and J. C. S. {Jacques Junior} and M. {Madadi} and S. {Ayache} and E. {Viegas} and F. {Gurpinar} and A. S. {Wicaksana} and C. {Liem} and M. A. J. {Van Gerven} and R. {Van Lier}},
  journal={IEEE Transactions on Affective Computing},
  title={Modeling, Recognizing, and Explaining Apparent Personality from Videos},
  year={2020},
  doi={10.1109/TAFFC.2020.2973984}
}

#########################################################
DISCLAIMER

Despite all of the efforts devoted to the compilation and curation of resources available in this website, we cannot guarantee that collected data including associated annotations and labels (obtained through manually, automatically and a semi-automatically processes) are representative samples of a real application scenario. The adopted data gathering and labeling methodologies may not include exhaustive and/or inclusive mechanisms that allow users to reach conclusive findings. More importantly, we strongly advise users NOT using the resources available in this site to build systems that make decisions and recommendations that have a direct or indirect impact into people's lives. Likewise, we acknowledge and apologize for those resources and publications available in this site may use ambiguous terms (e.g., gender vs. sex or ethnicity vs. race), we have not deliberately aimed to cause controversy or affect users in any form.
#########################################################