This project addresses limitations of the conventional surface EMG (sEMG) such as low sampling rates and a small number of EMG channels by developing a high-density sEMG (HD-sEMG) sensor and an machine learning (ML) algorithm to predict hand gestures from HD-sEMG. The developed HD-sEMG sensor includes 1,024 channels with a 20-kHz sampling rate (Fig. 1A). Junseong compares various ML algorithms to predict 6 different gestures (Fig. 1B). As HD-sEMG provides high-quality signals with a good discriminative ability (Fig. 1C), most algorithms can predict gestures near perfectly, even including the simplest linear discriminant analysis (LDA) method.