Tag
Introduces a montage-agnostic encoder for calibration-light cross-user gesture recognition from surface EMG, using shared weights and electrode coordinates to handle variable channel counts and reduce per-user calibration. It outperforms per-user baselines on some datasets and analyzes factors affecting cross-user transfer.
This paper introduces EMG-CrossFormer, a hybrid convolutional-transformer model for multimodal sEMG hand gesture recognition, which achieves state-of-the-art accuracy on NinaPro datasets by fusing sEMG and inertial signals via cross-attention.