Exploring Large-scale Unlabeled Faces to Enhance Facial Expression Recognition

TL;DR

This paper introduces a semi-supervised learning framework leveraging large-scale unlabeled face data with a dynamic threshold module to improve facial expression recognition performance, demonstrating effectiveness on the ABAW5 EXPR task.

Contribution

The paper presents a novel semi-supervised approach with an adaptive threshold module to utilize unlabeled face data for enhancing FER models.

Findings

Achieved excellent results on ABAW5 validation set

Effectively utilizes unlabeled face data for FER

Improves model generalization in facial expression recognition

Abstract

Facial Expression Recognition (FER) is an important task in computer vision and has wide applications in human-computer interaction, intelligent security, emotion analysis, and other fields. However, the limited size of FER datasets limits the generalization ability of expression recognition models, resulting in ineffective model performance. To address this problem, we propose a semi-supervised learning framework that utilizes unlabeled face data to train expression recognition models effectively. Our method uses a dynamic threshold module (\textbf{DTM}) that can adaptively adjust the confidence threshold to fully utilize the face recognition (FR) data to generate pseudo-labels, thus improving the model's ability to model facial expressions. In the ABAW5 EXPR task, our method achieved excellent results on the official validation set.