M2FNet: Multi-modal Fusion Network for Emotion Recognition in Conversation

TL;DR

This paper introduces M2FNet, a multi-modal fusion network that combines audio, visual, and text data using attention mechanisms and a novel loss function, significantly improving emotion recognition accuracy across multiple datasets.

Contribution

The paper presents a new multi-modal fusion architecture with a specialized feature extractor and adaptive triplet loss, advancing emotion recognition in conversations.

Findings

Outperforms existing methods on MELD and IEMOCAP datasets

Achieves state-of-the-art weighted F1 scores in ERC

Effectively integrates multi-modal data for emotion recognition

Abstract

Emotion Recognition in Conversations (ERC) is crucial in developing sympathetic human-machine interaction. In conversational videos, emotion can be present in multiple modalities, i.e., audio, video, and transcript. However, due to the inherent characteristics of these modalities, multi-modal ERC has always been considered a challenging undertaking. Existing ERC research focuses mainly on using text information in a discussion, ignoring the other two modalities. We anticipate that emotion recognition accuracy can be improved by employing a multi-modal approach. Thus, in this study, we propose a Multi-modal Fusion Network (M2FNet) that extracts emotion-relevant features from visual, audio, and text modality. It employs a multi-head attention-based fusion mechanism to combine emotion-rich latent representations of the input data. We introduce a new feature extractor to extract latent features from the audio and visual modality. The proposed feature extractor is trained with a novel adaptive margin-based triplet loss function to learn emotion-relevant features from the audio and visual data. In the domain of ERC, the existing methods perform well on one benchmark dataset but not on others. Our results show that the proposed M2FNet architecture outperforms all other methods in terms of weighted average F1 score on well-known MELD and IEMOCAP datasets and sets a new state-of-the-art performance in ERC.