MFC-wild (Many Faces of Confusion in the wild) is a dataset of short video clips (1-3s) from YouTube.com and Giphy.com using search tags such as "confusion", "anger" and "disgust" reactions to gather human facial expressions.
Type of data:
Video data, in mp4 file format.
Data is divided into folds for easily running the experiments explained in the paper (5-fold cross-validation, focusing on having one or two ethnic folds for each emotional class)
Our data is publicly available for research purposes. Please cite our paper (Accepted to IROS2022):
Akhyani et al., Using Sim2Real to Address Underrepresentation in Emotion Expression Recognition for Human-Robot Interaction, 2022
@INPROCEEDINGS{Akhyani2022,
author={Akhyani, Saba and Abbasi Boroujeni, Mehryar and Chen, Mo and Lim, Angelica},
booktitle={2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)(IROS 2022)},
title={ Towards Inclusive HRI: Using Sim2Real to Address Underrepresentation in Emotion Expression Recognition}},
year={2022}
}