MultiMAE-DER: Multimodal Masked Autoencoder for Dynamic Emotion Recognition

TL;DR

This paper introduces MultiMAE-DER, a multimodal masked autoencoder that improves dynamic emotion recognition by leveraging cross-modal correlations and optimizing fusion strategies, achieving state-of-the-art results on multiple datasets.

Contribution

The paper proposes a novel multimodal masked autoencoder approach with optimized fusion strategies for dynamic emotion recognition, outperforming existing models in supervised and self-supervised settings.

Findings

WAR improved by 4.41% on RAVDESS

WAR improved by 2.06% on CREMAD

WAR improved by 1.86% on IEMOCAP

Abstract

This paper presents a novel approach to processing multimodal data for dynamic emotion recognition, named as the Multimodal Masked Autoencoder for Dynamic Emotion Recognition (MultiMAE-DER). The MultiMAE-DER leverages the closely correlated representation information within spatiotemporal sequences across visual and audio modalities. By utilizing a pre-trained masked autoencoder model, the MultiMAEDER is accomplished through simple, straightforward finetuning. The performance of the MultiMAE-DER is enhanced by optimizing six fusion strategies for multimodal input sequences. These strategies address dynamic feature correlations within cross-domain data across spatial, temporal, and spatiotemporal sequences. In comparison to state-of-the-art multimodal supervised learning models for dynamic emotion recognition, MultiMAE-DER enhances the weighted average recall (WAR) by 4.41% on the RAVDESS dataset and by 2.06% on the CREMAD. Furthermore, when compared with the state-of-the-art model of multimodal self-supervised learning, MultiMAE-DER achieves a 1.86% higher WAR on the IEMOCAP dataset.