A Multi-System Acoustic Myography Dataset of Forearm Muscle Activity for Hand Gesture and Subject Recognition

This repository provides three anonymized acoustic myography (AMG) datasets of forearm muscle activity acquired with three different sensor systems. The datasets support research on hand gesture recognition, subject recognition, inter-subject variability, and the comparison of AMG acquisition systems. Dataset01 contains eight-channel AMG recordings from 20 participants performing 14 hand movements at 1024 Hz. Dataset02 contains six-channel recordings from 15 participants performing 11 movements at 1000 Hz, with multiple sessions available for some participants. Dataset03 contains six-channel recordings from 10 participants performing 8 movements at 1024 Hz. The AMG signals represent relative acoustic pressure variations expressed in Pascals (Pa). The data are provided as MATLAB files together with anonymous participant identifiers, session identifiers, harmonized movement labels, sensor descriptions, channel information, and machine-readable metadata. Original participant identifiers, correspondence tables, and private conversion files are not included in the public archive. Detailed information about the datasets, acquisition systems, signal format, movement classes, metadata files, and anonymization procedure is provided in DATASET_DESCRIPTION.md, included with this record. The data and metadata are released under the Creative Commons Attribution 4.0 International License (CC BY 4.0).