Electromyography and Inertial Data for Hand Gesture Recognition

A dataset containing electromyography and inertial measurement data for 14 distinct hand gestures, including common movements, finger gestures, and functional grips. Data were collected using Delsys Trigno sensors from participants performing gestures with their non-dominant hand, with sensors placed in both anatomical and random configurations on the forearm. The dataset was created by Yiling Zhang and last updated in May 2026.

Use Cases

• Training gesture classification models based on sEMG and IMU signals.
• Comparing the performance of anatomical versus random sensor placements for myoelectric control.
• Developing robust hand gesture recognition systems for assistive technologies.
• Analyzing the biomechanical signatures of functional grips like holding a mug or pen.

Strengths

• Includes data from two distinct sensor placement modes (anatomical and random) for comparative analysis.
• Covers 14 distinct hand gestures, including functional grips, performed ten times each.
• Uses high-frequency sampling (up to 4000 Hz for sEMG) with specified sensor ranges and equipment details.
• Ethical oversight was provided by the University Medical Center Groningen, with participant consent obtained.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• For 32 participants, the order of sensors in the random placement mode is unknown due to a technical issue.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

DataverseNL Harvested Dataverse, author Yiling Zhang.

Collection Method

Recorded using Delsys Trigno Discover system with Avanti and Quattro sensors, and annotated using Azure Kinect video.

Freshness

Last updated 2026-05-18 06:13:19; freshness should be verified.