Transformer-Based Self-Supervised Multimodal Representation Learning for Wearable Emotion Recognition
Open accessThe DOI matches its Crossref record; 13 of the 13 references it cites were found in a registry or an index.
Abstract
Recently, wearable emotion recognition based on peripheral physiological signals has drawn massive attention due to its less invasive nature and its applicability in real-life scenarios. However, how to effectively fuse multimodal data remains a challenging problem. Moreover, traditional fully-supervised based approaches suffer from overfitting given limited labeled data. To address the above issues, we propose a novel self-supervised learning (SSL) framework for wearable emotion recognition, where efficient multimodal fusion is realized with temporal convolution-based modality-specific encoders and a transformer-based shared encoder, capturing both intra-modal and inter-modal correlations. Extensive unlabeled data is automatically assigned labels by five signal transforms, and the proposed SSL model is pre-trained with signal transformation recognition as a pretext task, allowing the extraction of generalized multimodal representations for emotion-related downstream tasks. For evaluation, the proposed SSL model was first pre-trained on a large-scale self-collected physiological dataset and the resulting encoder was subsequently frozen or fine-tuned on three public supervised emotion recognition datasets. Ultimately, our SSL-based method achieved state-of-the-art results in various emotion classification tasks. Meanwhile, the proposed model was proved to be more accurate and robust compared to fully-supervised methods on low data regimes.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173
Hannun, A. Y., Rajpurkar, P., Haghpanahi, M., Tison, G. H., Bourn, C., Turakhia, M. P., & Ng, A. Y. (2019) · Nature Medicine · doi:10.1038/s41591-018-0268-3 · Cited by 3,148
Multimodal Transformer for Unaligned Multimodal Language Sequences
Tsai, Y.-H. H., Bai, S., Liang, P. P., Kolter, J. Z., Morency, L.-P., & Salakhutdinov, R. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/p19-1656 · Cited by 2,259
Detecting Stress During Real-World Driving Tasks Using Physiological Sensors
Healey, J. A., & Picard, R. W. (2005) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2005.848368 · Cited by 2,207
wav2vec: Unsupervised Pre-Training for Speech Recognition
Schneider, S., Baevski, A., Collobert, R., & Auli, M. (2019) · Interspeech 2019 · doi:10.21437/interspeech.2019-1873 · Cited by 1,290
Um, T. T., Pfister, F. M. J., Pichler, D., Endo, S., Lang, M., Hirche, S., Fietzek, U., & Kulić, D. (2017) · Proceedings of the 19th ACM International Conference on Multimodal Interaction · doi:10.1145/3136755.3136817 · Cited by 729 · cited 1×
“Meanwhile, according to the review of time series augmentation strategies [30], [31], though most of the suggested transformations have been adopted in previous SSL-based work, two transformations have not been thoroughly evaluated: Magnitudewarping and Cropping.”
BackgroundAn empirical survey of data augmentation for time series classification with neural networks
Iwana, B. K., & Uchida, S. (2021) · PLOS ONE · doi:10.1371/journal.pone.0254841 · Cited by 725
Multi-task Self-Supervised Learning for Human Activity Detection
Saeed, A., Ozcelebi, T., & Lukkien, J. (2019) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3328932 · Cited by 346 · cited 5×
“The followings are brief descriptions of these models: Supervised learning-based methods: (cid:2) SimpDCNN [27]: it is a simple convolutional network consisting of three convolutional blocks with kernel sizes of 24, 16 and 8, each followed by a ReLU activation and a dropout layer.”
MethodResultInfluentialSantamaria-Granados, L., Munoz-Organero, M., Ramirez-Gonzalez, G., Abdulhay, E., & Arunkumar, N. (2019) · IEEE Access · doi:10.1109/access.2018.2883213 · Cited by 289 · cited 3×
“Joint representations can be extracted directly from concatenated vectors with deep models such as 1DCNN [12] and Bi-LSTM [21], which allow for encoding inter-modal correlations.”
BackgroundMethodResultEmotion Recognition using Multimodal Residual LSTM Network
Ma, J., Tang, H., Zheng, W.-L., & Lu, B.-L. (2019) · Proceedings of the 27th ACM International Conference on Multimedia · doi:10.1145/3343031.3350871 · Cited by 230 · cited 1×
“(cid:2) MMResLSTM [38]: it uses separate four-layer LSTM-based models for multimodal signals with residual connections.”
MethodEmotion Recognition from Multimodal Physiological Signals for Emotion Aware Healthcare Systems
Ayata, D., Yaslan, Y., & Kamasak, M. E. (2020) · Journal of Medical and Biological Engineering · doi:10.1007/s40846-019-00505-7 · Cited by 182 · cited 2×
“Late fusion-based approaches [2], [14] integrate the decisions of multiple independent learning models to predict emotion categories.”
BackgroundSubject-Aware Contrastive Learning for Biosignals
Cheng, J. Y., Goh, H., Dogrusoz, K., Tuzel, O., & Azemi, E. (2020) · arXiv · doi:10.48550/arxiv.2007.04871 · Cited by 74 · cited 1×
“Several SSL-based efforts [8], [9], [10] have been done for emotion recognition using EEG signal, but they are not suitable for practical scenes.”
Background
Cited by
A systematic review of intermediate fusion in multimodal deep learning for biomedical applications
Guarrasi, V., Aksu, F., Caruso, C. M., Di Feola, F., Rofena, A., Ruffini, F., & Soda, P. (2025) · Image and Vision Computing · doi:10.1016/j.imavis.2025.105509 · Cited by 95
State-of-the-art Multimodal Emotion Recognition: A comprehensive survey and taxonomy
Yazici, A., Kucukyilmaz, T., Dokeroglu, T., Sharipbay, A., Lee, M.-H., & Tyler, B. (2026) · Intelligent Systems with Applications · doi:10.1016/j.iswa.2026.200642 · Cited by 11
Alam, M. S. B., Lameesa, A., Sharmin, S., Afrin, S., Ahmed, S. F., Nikoo, M. R., & Gandomi, A. H. (2025) · Digital Communications and Networks · doi:10.1016/j.dcan.2025.04.001 · Cited by 9
Metadata
- Authors
- Wu, Y., Daoudi, M., & Amad, A.
- Venue
- IEEE Transactions on Affective Computing
- Published
- 2024-01
- Volume
- 15
- Issue
- 1
- Pages
- 157-172
- ISSN
- 1949-3045 (ISSN-L)
- 2371-9850
- OpenAlex
- W4362496488
- Cited by
- 106
- References
- 13
- FWCI
- 24.61
- Type
- journal article
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.