TUM-ICS/EMG-soft-hand-exoskeleton: v1.0.0 — Initial Release

v1.0.0 — Initial Release This release accompanies the scientific publication on EMG-based activity onset detection for hand exoskeleton control in persons with hand impairment (ALS, SMA) and healthy individuals. Contents 1D CNN models for activity onset detection on RMS-processed EMG signals (34.81 Hz) Adaptive threshold onset detector for comparison Training and evaluation scripts, including the main replication script 1_evaluate_cnn_onset_detection.py Grad-CAM analysis tools for model interpretability EMG signal data and ground-truth onset labels for 15 healthy participants and ALS/SMA patient recordings Reproducibility Python 3.12.1, dependencies pinned via uv.lock / requirements.txt Step-by-step replication instructions in the README.md Citation If you use this code or data, please cite the associated publication (see README.md).