EUVIMEDEUVIMED entdecken
Crossref

Adaptive PPG Decomposition and Skin Temperature Guidance for Wearable Emotion Recognition

Lingyun Zhu, Guojie Sun, Daohong Wei, Dongyi Chen

IEEE Signal Processing Letters · 2026 · 33 · 4221-4225

DOI: 10.1109/lsp.2026.3738658 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Lingyun Zhu und 3 weiteren Autor:innen. Die Arbeit nennt 38 Literaturangaben. Gefördert von: National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Photoplethysmography (PPG) offers low power consumption and minimal intrusiveness, making it inherently advantaged for continuous, long-term wearable emotion monitoring. Nevertheless, PPG-based emotion recognition still faces persistent challenges, including scarce labeled data, poor cross-subject generalization due to substantial physiological variability, and multi-component aliasing inherent to physiological signals. Moreover, existing multimodal methods treat modalities as parallel inputs, failing to exploit cross-modal physiological synergy. This paper proposes a dual-stage, dynamically calibrated multimodal framework that achieves efficient cross-modal synergy via self-supervised pre-training and skin temperature (SKT) guided calibration. In the pre-training stage, a Pulse-Decomposed Mamba encoder (PD-Mamba) is embedded within a Masked Autoencoder (MAE) framework. Through masked reconstruction on large-scale unlabeled PPG data, the model learns generalizable physiological representations. In the downstream stage, SKT serves as an explicit physiological prior. A Skin Temperature-Guided Attention (STGA) module dynamically calibrates PPG features, while a Symplectic Dynamic Fusion (SDF) network enables adaptive deep PPG–SKT fusion. Extensive subject-independent evaluations across multiple datasets confirm the framework's high accuracy and robust generalization.

Förderung

  • National Natural Science Foundation of China

Zitieren

APA

Zhu, L., Sun, G., Wei, D., & Chen, D. (2026). Adaptive PPG Decomposition and Skin Temperature Guidance for Wearable Emotion Recognition. IEEE Signal Processing Letters, 33, 4221-4225. https://doi.org/10.1109/lsp.2026.3738658

Vancouver

Zhu L, Sun G, Wei D, Chen D. Adaptive PPG Decomposition and Skin Temperature Guidance for Wearable Emotion Recognition. IEEE Signal Processing Letters. 2026;33:4221-4225. doi:10.1109/lsp.2026.3738658

BibTeX
@article{Zhu2026,
  author = {Zhu, Lingyun and Sun, Guojie and Wei, Daohong and Chen, Dongyi},
  title = {Adaptive PPG Decomposition and Skin Temperature Guidance for Wearable Emotion Recognition},
  journal = {IEEE Signal Processing Letters},
  year = {2026},
  volume = {33},
  pages = {4221--4225},
  doi = {10.1109/lsp.2026.3738658}
}
Kostenlos starten →