A Peek at Peak Emotion Recognition

TL;DR

This paper investigates deep learning's ability to recognize peak emotions in facial expressions, showing it outperforms humans even with limited data and human-annotated datasets.

Contribution

It demonstrates that deep learning models can surpass human performance in peak emotion recognition despite small datasets and reliance on human labels.

Findings

Deep learning models outperform humans in peak emotion recognition.

Models trained on small datasets still achieve high accuracy.

Deep features outperform human judgment in emotion discernment.

Abstract

Despite much progress in the field of facial expression recognition, little attention has been paid to the recognition of peak emotion. Aviezer et al. [1] showed that humans have trouble discerning between positive and negative peak emotions. In this work we analyze how deep learning fares on this challenge. We find that (i) despite using very small datasets, features extracted from deep learning models can achieve results significantly better than humans. (ii) We find that deep learning models, even when trained only on datasets tagged by humans, still outperform humans in this task.