What Does it Take to Generalize SER Model Across Datasets? A Comprehensive Benchmark

TL;DR

This paper presents a comprehensive benchmark for speech emotion recognition (SER) across multiple datasets, addressing generalization challenges and evaluating the potential of Whisper, with a focus on speaker-independent methods and data imbalance solutions.

Contribution

It introduces a large-scale benchmark with 11 datasets, explores data balancing and evaluation protocols, and assesses Whisper's effectiveness for SER.

Findings

Improved generalization across diverse datasets

Effective handling of data imbalance with over-sampling

Whisper shows promising results for SER

Abstract

Speech emotion recognition (SER) is essential for enhancing human-computer interaction in speech-based applications. Despite improvements in specific emotional datasets, there is still a research gap in SER's capability to generalize across real-world situations. In this paper, we investigate approaches to generalize the SER system across different emotion datasets. In particular, we incorporate 11 emotional speech datasets and illustrate a comprehensive benchmark on the SER task. We also address the challenge of imbalanced data distribution using over-sampling methods when combining SER datasets for training. Furthermore, we explore various evaluation protocols for adeptness in the generalization of SER. Building on this, we explore the potential of Whisper for SER, emphasizing the importance of thorough evaluation. Our approach is designed to advance SER technology by integrating speaker-independent methods.