Back to datasets
Dataset assetOpen Source CommunityHuman Motion AnalysisElectromyography

myo-readings-dataset

The Myo‑armband EMG reading dataset contains recordings for multiple wrist gestures, including rest, flexion, extension, radial deviation, ulnar deviation, pronation, supination, and fist. The dataset details the data structure, recording protocol, gesture labels, and file formats.

Source
github
Created
Feb 5, 2021
Updated
Apr 6, 2024
Signals
297 views
Availability
Linked source ready
Overview

Dataset description and usage context

myo‑readings‑dataset Overview

Dataset Description

This dataset contains EMG readings captured by a Myo armband to record wrist actions such as rest, flexion, extension, radial deviation, ulnar deviation, pronation, supination, and fist.

Project Structure

  • Right‑hand readings: located in the _readings_right_hand folder.
  • Left‑hand readings: located in the _readings_left_hand folder.
  • Participant data: each folder is named with a random unique five‑digit ID followed by a session number, e.g., 12345-1.
  • Session folder: contains multiple files, each corresponding to a wrist gesture; filenames are <label>.txt, e.g., 2.txt denotes extension.
  • Each session folder should contain at least eight files for gestures 0‑7.
  • Each file: consists of many lines; each line records the eight EMG channel values from the Myo armband and the wrist‑gesture label, formatted as <EMG>,<EMG>,...,<label>.

Recording Protocol

  • Myo placement: on the thickest part of the forearm with the LED facing the dorsal side of the hand.
  • Each file: contains one minute of recording (~12,000 lines), each line representing a time point.
  • Gesture labeling: the label changes every five lines; rest‑gesture files contain only the rest label (0).

Gesture Labels

  • 0: Rest
  • 1: Flexion
  • 2: Extension
  • 3: Radial deviation
  • 4: Ulnar deviation
  • 5: Pronation
  • 6: Supination
  • 7: Fist

Curated Sessions

The curated.txt file lists sessions that exhibit good intra‑participant accuracy across most gestures. It is recommended to use only these curated sessions when applying machine‑learning strategies to the dataset.

Need downstream help?

Pair the dataset with AI analysis and content workflows.

Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.

Explore AI studio