Pairwise Discernment of AffectNet Expressions with ArcFace

TL;DR

This paper explores using transfer learning with ArcFace and other models on AffectNet to improve facial emotion recognition, emphasizing pairwise learning to handle class imbalance and enhance accuracy.

Contribution

It demonstrates the effectiveness of domain-specific transfer learning and pairwise learning techniques in improving FER performance on imbalanced datasets.

Findings

Transfer learning with ArcFace improves FER accuracy.

Pairwise learning helps address class imbalance.

Domain-specific transfer learning is valuable for FER.

Abstract

This study takes a preliminary step toward teaching computers to recognize human emotions through Facial Emotion Recognition (FER). Transfer learning is applied using ResNeXt, EfficientNet models, and an ArcFace model originally trained on the facial verification task, leveraging the AffectNet database, a collection of human face images annotated with corresponding emotions. The findings highlight the value of congruent domain transfer learning, the challenges posed by imbalanced datasets in learning facial emotion patterns, and the effectiveness of pairwise learning in addressing class imbalances to enhance model performance on the FER task.