@article{ART003338866},
author={Duyong Baek and Seok-Won Lee},
title={EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition},
journal={Journal of The Korea Society of Computer and Information},
issn={1598-849X},
year={2026},
volume={31},
number={5},
pages={41-53}
TY - JOUR
AU - Duyong Baek
AU - Seok-Won Lee
TI - EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition
JO - Journal of The Korea Society of Computer and Information
PY - 2026
VL - 31
IS - 5
PB - The Korean Society Of Computer And Information
SP - 41
EP - 53
SN - 1598-849X
AB - High-density electroencephalography (EEG) systems provide superior spatial resolution for emotion recognition, but their high cost, complexity, and user discomfort make them impractical for real-world applications. This study presents a practical alternative: an EEG super-resolution paradigm operating in the Power Spectral Density (PSD) feature space rather than the raw signal domain. The proposed Transformer-based architecture reconstructs 62-channel high-density PSD features from sparse inputs of 8, 14, 16, and 32 channels through a subject-independent reconstruction model. Evaluated on the SEED-IV dataset, the proposed method achieves an average Accuracy Preservation Ratio (APR) of 78.45% with 32 channels and 77.92% with 16 channels (Uniform strategy), reaching up to 97.56%. The Uniform channel selection strategy significantly outperforms both Clinical Standard and Hardware Realistic strategies (paired t-test, p<0.01, Bonferroni-corrected). These results demonstrate that low-channel wearable EEG can achieve emotion recognition performance comparable to clinical-grade systems without subject-specific high-density data collection.
KW - EEG Super-resolution;Emotion Recognition;Power Spectral Density;Transformer;Wearable EEG;Subject-independent Reconstruction
DO -
UR -
ER -
Duyong Baek and Seok-Won Lee. (2026). EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition. Journal of The Korea Society of Computer and Information, 31(5), 41-53.
Duyong Baek and Seok-Won Lee. 2026, "EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition", Journal of The Korea Society of Computer and Information, vol.31, no.5 pp.41-53.
Duyong Baek, Seok-Won Lee "EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition" Journal of The Korea Society of Computer and Information 31.5 pp.41-53 (2026) : 41.
Duyong Baek, Seok-Won Lee. EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition. 2026; 31(5), 41-53.
Duyong Baek and Seok-Won Lee. "EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition" Journal of The Korea Society of Computer and Information 31, no.5 (2026) : 41-53.
Duyong Baek; Seok-Won Lee. EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition. Journal of The Korea Society of Computer and Information, 31(5), 41-53.
Duyong Baek; Seok-Won Lee. EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition. Journal of The Korea Society of Computer and Information. 2026; 31(5) 41-53.
Duyong Baek, Seok-Won Lee. EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition. 2026; 31(5), 41-53.
Duyong Baek and Seok-Won Lee. "EEG Super-Resolution in PSD Feature Space: Bridging Wearable and Clinical-Grade Emotion Recognition" Journal of The Korea Society of Computer and Information 31, no.5 (2026) : 41-53.
