Enhancing Speech Emotion Recognition through Segmental Average Pooling of Self-Supervised Learning Features
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910652707110912 |
|---|---|
| author | Hyeon, Jonghwan Oh, Yung-Hwan Choi, Ho-Jin |
| author_facet | Hyeon, Jonghwan Oh, Yung-Hwan Choi, Ho-Jin |
| contents | Speech Emotion Recognition (SER) analyzes human emotions expressed through speech. Self-supervised learning (SSL) offers a promising approach to SER by learning meaningful representations from a large amount of unlabeled audio data. However, existing SSL-based methods rely on Global Average Pooling (GAP) to represent audio signals, treating speech and non-speech segments equally. This can lead to dilution of informative speech features by irrelevant non-speech information. To address this, the paper proposes Segmental Average Pooling (SAP), which selectively focuses on informative speech segments while ignoring non-speech segments. By applying both GAP and SAP to SSL features, our approach utilizes overall speech signal information from GAP and specific information from SAP, leading to improved SER performance. Experiments show state-of-the-art results on the IEMOCAP for English and superior performance on KEMDy19 for Korean datasets in both unweighted and weighted accuracies. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_12416 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Enhancing Speech Emotion Recognition through Segmental Average Pooling of Self-Supervised Learning Features Hyeon, Jonghwan Oh, Yung-Hwan Choi, Ho-Jin Sound Artificial Intelligence Audio and Speech Processing Speech Emotion Recognition (SER) analyzes human emotions expressed through speech. Self-supervised learning (SSL) offers a promising approach to SER by learning meaningful representations from a large amount of unlabeled audio data. However, existing SSL-based methods rely on Global Average Pooling (GAP) to represent audio signals, treating speech and non-speech segments equally. This can lead to dilution of informative speech features by irrelevant non-speech information. To address this, the paper proposes Segmental Average Pooling (SAP), which selectively focuses on informative speech segments while ignoring non-speech segments. By applying both GAP and SAP to SSL features, our approach utilizes overall speech signal information from GAP and specific information from SAP, leading to improved SER performance. Experiments show state-of-the-art results on the IEMOCAP for English and superior performance on KEMDy19 for Korean datasets in both unweighted and weighted accuracies. |
| title | Enhancing Speech Emotion Recognition through Segmental Average Pooling of Self-Supervised Learning Features |
| topic | Sound Artificial Intelligence Audio and Speech Processing |
| url | https://arxiv.org/abs/2410.12416 |