DWFormer: Dynamic Window transFormer for Speech Emotion Recognition

TL;DR

DWFormer is a novel transformer-based model that dynamically segments speech into windows to better capture emotion-related temporal features at multiple scales, improving speech emotion recognition accuracy.

Contribution

It introduces a dynamic windowing mechanism within a transformer architecture to enhance local and global temporal feature extraction for speech emotion recognition.

Findings

Outperforms previous state-of-the-art methods on IEMOCAP and MELD datasets.

Effectively captures multi-scale temporal information for emotion recognition.

Demonstrates improved accuracy over existing models.

Abstract

Speech emotion recognition is crucial to human-computer interaction. The temporal regions that represent different emotions scatter in different parts of the speech locally. Moreover, the temporal scales of important information may vary over a large range within and across speech segments. Although transformer-based models have made progress in this field, the existing models could not precisely locate important regions at different temporal scales. To address the issue, we propose Dynamic Window transFormer (DWFormer), a new architecture that leverages temporal importance by dynamically splitting samples into windows. Self-attention mechanism is applied within windows for capturing temporal important information locally in a fine-grained way. Cross-window information interaction is also taken into account for global communication. DWFormer is evaluated on both the IEMOCAP and the MELD datasets. Experimental results show that the proposed model achieves better performance than the previous state-of-the-art methods.