Author name cluster

Zepeng Yu

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

2 papers

1 author row

AAAI Conference 2026 Conference Paper

CAT-Net: A Cross-Attention Tone Network for Cross-Subject EEG-EMG Fusion Tone Decoding

Yifan Zhuang
Calvin Huang
Zepeng Yu
Yongjie Zou
Jiawei Ju

Brain-computer interface (BCI) speech decoding has emerged as a promising tool for assisting individuals with speech impairments. In this context, the integration of electroencephalography (EEG) and electromyography (EMG) signals offers strong potential for enhancing decoding performance. Mandarin tone classification presents particular challenges, as tonal variations convey distinct meanings even when phonemes remain identical. In this study, we propose a novel cross-subject multimodal BCI decoding framework that fuses EEG and EMG signals to classify four Mandarin tones under both audible and silent speech conditions. Inspired by the cooperative mechanisms of neural and muscular systems in speech production, our neural decoding architecture combines spatial-temporal feature extraction branches with a cross-attention fusion mechanism, enabling informative interaction between modalities. We further incorporate domain-adversarial training to improve cross-subject generalization. We collected 4,800 EEG trials and 4,800 EMG trials from 10 participants using only twenty EEG and five EMG channels, demonstrating the feasibility of minimal-channel decoding. Despite employing lightweight modules, our model outperforms state-of-the-art baselines across all conditions, achieving average classification accuracies of 87.83\% for audible speech and 88.08\% for silent speech. In cross-subject evaluations, it still maintains strong performance with accuracies of 83.27\% and 85.10\% for audible and silent speech, respectively. We further conduct ablation studies to validate the effectiveness of each component. Our findings suggest that tone-level decoding with minimal EEG-EMG channels is feasible and potentially generalizable across subjects, contributing to the development of practical BCI applications.

PDF Details DOI

EAAI Journal 2024 Journal Article

Cross-modal misalignment-robust feature fusion for crowd counting

Weihang Kong
Zepeng Yu
He Li
Junge Zhang

Mainstream crowd counting methods in Red-Green-Blue(RGB)-Thermal(T) information processing field concentrate on how to realize cross-modal complementary feature fusion. However, the cross-modal image misalignment issue has almost not been concerned and discussed for the target task, which substantially affects the precise feature extraction and fusion. Given this, this work intends to mitigate the adverse effects of the misalignment issue between the visible and thermal modalities on complementary feature representation. Specially, we design a cross-modal feature alignment fusion network for crowd counting, with a cross-modal feature alignment (CFA) module and triple-branch frequency-cascaded fusion (TFF) module as the core components. The CFA utilizes the design of the differential deformable representation to calibrate the cross-modal feature misalignment. Then the TFF employs three interactive differential branches with time-frequency joint feature modeling block to realize the mighty complementary feature representation. Extensive experiments, along with the ablation studies, show the effectiveness of the proposed method on the feature alignment and fusion over two challenging RGB-T crowd counting benchmarks. And the extended experimental results also suggest the feasibility of the proposed method on RGB-Depth crowd counting. Among them, the proposed method achieves 6. 97% improvement over the mean absolute error metrics than the suboptimal method on the surveillance-view benchmark, and 10. 72% improvement over the root mean square error. The development of the solution reveals the exploration over the cross-modal misalignment issue could promote the final counting effect, and also provides an effective resolution for the target RGB-Thermal crowd counting, as well as for other similar RGB-Thermal computer vision tasks.

Details DOI