Exploring Expression-related Self-supervised Learning for Affective Behaviour Analysis

TL;DR

This paper introduces ContraWarping, a self-supervised learning approach for expression classification that leverages unlabeled data, outperforming many supervised methods in affective behavior analysis.

Contribution

The paper presents a novel SSL method, ContraWarping, specifically designed for expression classification in affective behavior analysis, demonstrating superior performance on the Aff-Wild2 dataset.

Findings

ContraWarping outperforms most existing supervised methods

Effective use of unlabeled data in affective analysis

Potential for broad application in affective behavior tasks

Abstract

This paper explores an expression-related self-supervised learning (SSL) method (ContraWarping) to perform expression classification in the 5th Affective Behavior Analysis in-the-wild (ABAW) competition. Affective datasets are expensive to annotate, and SSL methods could learn from large-scale unlabeled data, which is more suitable for this task. By evaluating on the Aff-Wild2 dataset, we demonstrate that ContraWarping outperforms most existing supervised methods and shows great application potential in the affective analysis area. Codes will be released on: https://github.com/youqingxiaozhua/ABAW5.