FaceExpressions-70k: A Dataset of Perceived Expression Differences

Accepted at ACM SIGGRAPH 2025

Avinab Saha1, Yu-Chih Chen1, Jean-Charles Bazin2, Christian Häne2, Ioannis Katsavounidis2, Alexandre Chapiro2, Alan C. Bovik1
1 The University of Texas at Austin UT Austin Logo
2 Reality Labs, Meta Meta Logo
Paper (PDF) Paper (HTML) Dataset YouTube Video BibTeX
Teaser image for FaceExpressions-70k
Sample frames from the FaceExpressions-70k dataset. The images are sampled from expression transition videos in the MultiFace Dataset [Wuu et al. 2022]. A large-scale human study on Amazon Mechanical Turk uses the sampled image pairs to evaluate perceptual expression differences. Intra-expression pairs from the same expression sequences have expression difference scores in the top right triangles. Inter-expression pairs, comparing across expressions, are shown in the matrix in the bottom right. Black cells with white crosses indicate excluded inter-expression pairs to manage the scale of the study.

Abstract

Facial expressions are key to human communication, conveying emotions and intentions. Given the rising popularity of digital humans and avatars, the ability to accurately represent facial expressions in real time has become an important topic. However, quantifying perceived differences between pairs of expressions is difficult, and no comprehensive subjective datasets are available for testing. This work introduces a new dataset targeting this problem: FaceExpressions-70k. Obtained via crowdsourcing, our dataset contains 70,500 subjective expression comparisons rated by over 1,000 study participants. We demonstrate the applicability of the dataset for training perceptual expression difference models and guiding decisions on acceptable latency and sampling rates for facial expressions when driving a face avatar.

Citation

If you use our dataset in your work, please cite:

@inproceedings{saha2025faceexpressions,
  author    = {Saha, Avinab and Chen, Yu-Chih and Bazin, Jean-Charles and H{\"a}ne, Christian and Katsavounidis, Ioannis and Chapiro, Alexandre and Bovik, Alan C},
  title     = {FaceExpressions-70k: A Dataset of Perceived Expression Differences},
  year      = {2025},
  publisher = {Association for Computing Machinery},
  address   = {New York, NY, USA},
  booktitle = {ACM SIGGRAPH 2025 Conference Proceedings},
  series    = {SIGGRAPH '25}
}