This study proposes a unified multimodal temporal motion state perception framework for optical imaging-oriented biomedical applications, integrating visual skeleton sequences, inertial measurement unit (IMU) signals, and surface electromyography (EMG) signals. The framework utilizes modality-specific encoders and a cross-modal temporal alignment attention mechanism to explicitly model temporal offsets from heterogeneous sensing streams. A multimodal temporal Transformer backbone is introduced to capture long-range motion dependencies and cross-modal interactions, while an uncertainty-aware fusion module dynamically allocates weights based on modality confidence. Experimental results demonstrate that the proposed approach achieves an accuracy of 94.37%, an F1-score of 93.95%, and a mean average precision of 96.02%, outperforming mainstream baseline models. Robustness evaluations further confirm stable performance under visual occlusion and sensor noise. These results indicate that the framework provides a highly accurate and robust solution for rehabilitation assessment, sports training monitoring, and wearable intelligent interaction systems.

Muscle onset detection is a fundamental problem in electromyography signal analysis, human–machine interaction, and rehabilitation assessment. In medical and biomedical applications, slow muscle activation onset processes are widely encountered in scenarios such as rehabilitation training, postural regulation, and fine motor control. Such processes are typically characterized by slowly varying amplitudes, long temporal durations, and high susceptibility to noise interference, which poses significant challenges for accurate identification of onset timing. To address these issues, a lightweight temporal attention method for slow muscle activation onset detection is proposed and systematically validated under multimodal experimental settings. The proposed method takes surface electromyography signals as the primary input, while synchronously acquired optical motion image data are incorporated into the experimental design and result analysis, thereby aligning with the common joint use of optical imaging and physiological signals in medical and biomedical research. From a methodological perspective, the proposed framework is composed of lightweight temporal feature encoding, a slow activation-aware temporal attention mechanism, and noise suppression with stable decision strategies. Under the constraint of low computational complexity, the ability to model progressive activation signals is effectively enhanced. Experiments are conducted on a dataset containing multiple types of slow activation movements, and model performance is evaluated using five-fold cross-validation. The results demonstrate that under regular signal-to-noise ratio conditions, the proposed method significantly outperforms traditional threshold-based approaches, classical machine learning models, and several deep learning baselines in terms of onset detection accuracy, recall, and precision. Specifically, onset detection accuracy reaches approximately $92\%$, recall is around $90\%$, and precision is approximately $93\%$. Meanwhile, the average onset detection error and detection delay are reduced to about $41\mathrm{ms}$ and $28\mathrm{ms}$, respectively, with the false positive rate controlled at approximately $2.2\%$. Stable performance is further maintained under different noise levels and cross-subject settings, indicating strong robustness and generalization capability.