Human Reaction Intensity Estimation with Ensemble of Multi-task Networks

TL;DR

This paper introduces a multi-task learning approach to estimate emotional reaction intensity from facial expressions in-the-wild, achieving promising preliminary results in a competitive benchmark.

Contribution

It presents a novel multi-emotional task learning framework specifically designed for emotional reaction intensity estimation in unconstrained environments.

Findings

Achieved a mean PCC score of 0.3254 on the ERI challenge.

Demonstrated the effectiveness of multi-task learning for ERI estimation.

Provided preliminary results for the ABAW competition.

Abstract

Facial expression in-the-wild is essential for various interactive computing domains. Especially, "Emotional Reaction Intensity" (ERI) is an important topic in the facial expression recognition task. In this paper, we propose a multi-emotional task learning-based approach and present preliminary results for the ERI challenge introduced in the 5th affective behavior analysis in-the-wild (ABAW) competition. Our method achieved the mean PCC score of 0.3254.