APS

2026 APS Annual Convention · 2026

Introducing Chehre: An Open-Access Dataset of Facial Expressions Based on Emojis

Barcelona, Spain · May 2026

Posters · How We Know: Methods, Measurement & Open Science

  • Bita Azari
    SFU
  • Zoe Stanley
    Simon Fraser University
  • Avneet Batra
    Simon Fraser University
  • Poorvi Bhatia
    Simon Fraser University
  • Manolis Savva
    Simon Fraser University
  • Hali Kil
    Simon Fraser University
  • Angelica Lim
    Simon Fraser University

Abstract

This exploratory study introduces Chehre, a multimodal emoji-expression dataset with data from n=843 participants. Dynamic facial-head portrayals of 40 emojis were transferred via neural face-swap onto generated photorealistic people. We contribute 456 high-quality videos with strong agreement, covering 37 emoji and 57 meaning labels, alongside semantic space theory analyses.

← Poster Session XI