Shuffling Augmented Decoupled Features for Multimodal Emotion Recognition

Multimodal emotion recognition (MER) aims to identify human emotions using data from multiple modalities. Despite promising advances in previous MER methods, their performance remains limited due to the small size of available datasets, a result of the challenges in collecting multimodal data. While...

Full description

Saved in:
Bibliographic Details
Main Author: Sunyoung Cho
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11014057/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850175107949920256
author Sunyoung Cho
author_facet Sunyoung Cho
author_sort Sunyoung Cho
collection DOAJ
description Multimodal emotion recognition (MER) aims to identify human emotions using data from multiple modalities. Despite promising advances in previous MER methods, their performance remains limited due to the small size of available datasets, a result of the challenges in collecting multimodal data. While data augmentation can address this issue, generating augmented multimodal data without altering the underlying emotional meaning remains particularly challenging. To tackle this problem, we introduce a decoupled feature augmentation method that automatically learns multimodal feature variations in a decoupled feature space for MER. Specifically, we decompose multimodal features into modality-invariant and modality-specific components and then augment each component within the decoupled feature space across multiple modalities. Unlike existing unimodal augmentation approaches, our method preserves cross-modal semantic consistency by jointly augmenting the decoupled components. To enhance model generalization and stability, we propose a learning strategy that gradually incorporates more diverse information by using a combined set of original and augmented decoupled features. Comprehensive experiments on two MER benchmarks demonstrate that our method outperforms or is comparable to several baseline methods.
format Article
id doaj-art-e2e9ea5cb3ac4a69b9bda6161ca429b9
institution OA Journals
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-e2e9ea5cb3ac4a69b9bda6161ca429b92025-08-20T02:19:31ZengIEEEIEEE Access2169-35362025-01-0113912909130010.1109/ACCESS.2025.357292511014057Shuffling Augmented Decoupled Features for Multimodal Emotion RecognitionSunyoung Cho0https://orcid.org/0000-0002-6925-6077Division of Software, Sookmyung Women’s University, Seoul, Republic of KoreaMultimodal emotion recognition (MER) aims to identify human emotions using data from multiple modalities. Despite promising advances in previous MER methods, their performance remains limited due to the small size of available datasets, a result of the challenges in collecting multimodal data. While data augmentation can address this issue, generating augmented multimodal data without altering the underlying emotional meaning remains particularly challenging. To tackle this problem, we introduce a decoupled feature augmentation method that automatically learns multimodal feature variations in a decoupled feature space for MER. Specifically, we decompose multimodal features into modality-invariant and modality-specific components and then augment each component within the decoupled feature space across multiple modalities. Unlike existing unimodal augmentation approaches, our method preserves cross-modal semantic consistency by jointly augmenting the decoupled components. To enhance model generalization and stability, we propose a learning strategy that gradually incorporates more diverse information by using a combined set of original and augmented decoupled features. Comprehensive experiments on two MER benchmarks demonstrate that our method outperforms or is comparable to several baseline methods.https://ieeexplore.ieee.org/document/11014057/Feature augmentationmultimodal emotion recognitionmultimodal learning
spellingShingle Sunyoung Cho
Shuffling Augmented Decoupled Features for Multimodal Emotion Recognition
IEEE Access
Feature augmentation
multimodal emotion recognition
multimodal learning
title Shuffling Augmented Decoupled Features for Multimodal Emotion Recognition
title_full Shuffling Augmented Decoupled Features for Multimodal Emotion Recognition
title_fullStr Shuffling Augmented Decoupled Features for Multimodal Emotion Recognition
title_full_unstemmed Shuffling Augmented Decoupled Features for Multimodal Emotion Recognition
title_short Shuffling Augmented Decoupled Features for Multimodal Emotion Recognition
title_sort shuffling augmented decoupled features for multimodal emotion recognition
topic Feature augmentation
multimodal emotion recognition
multimodal learning
url https://ieeexplore.ieee.org/document/11014057/
work_keys_str_mv AT sunyoungcho shufflingaugmenteddecoupledfeaturesformultimodalemotionrecognition