{"id":16203064,"url":"https://github.com/michidk/myo-dataset","last_synced_at":"2026-01-20T04:32:40.916Z","repository":{"id":73861288,"uuid":"350807043","full_name":"michidk/myo-dataset","owner":"michidk","description":"This repository contains sEMG Data of 13 subjects recorded with the Myo Armband.","archived":false,"fork":false,"pushed_at":"2022-01-08T09:36:39.000Z","size":18597,"stargazers_count":7,"open_issues_count":0,"forks_count":4,"subscribers_count":3,"default_branch":"main","last_synced_at":"2025-04-07T19:17:12.398Z","etag":null,"topics":["data-set","dataset","gesture-detection","gesture-recognition","myo","myo-armband","rock-paper-scissors","semg"],"latest_commit_sha":null,"homepage":"","language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/michidk.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2021-03-23T17:54:06.000Z","updated_at":"2024-12-15T23:19:50.000Z","dependencies_parsed_at":"2023-02-27T01:15:31.696Z","dependency_job_id":null,"html_url":"https://github.com/michidk/myo-dataset","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/michidk/myo-dataset","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/michidk%2Fmyo-dataset","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/michidk%2Fmyo-dataset/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/michidk%2Fmyo-dataset/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/michidk%2Fmyo-dataset/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/michidk","download_url":"https://codeload.github.com/michidk/myo-dataset/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/michidk%2Fmyo-dataset/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":28596079,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-01-20T02:08:49.799Z","status":"ssl_error","status_checked_at":"2026-01-20T02:08:44.148Z","response_time":117,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["data-set","dataset","gesture-detection","gesture-recognition","myo","myo-armband","rock-paper-scissors","semg"],"created_at":"2024-10-10T09:51:04.335Z","updated_at":"2026-01-20T04:32:40.899Z","avatar_url":"https://github.com/michidk.png","language":null,"funding_links":[],"categories":[],"sub_categories":[],"readme":"# Myo Dataset\n\nMyo dataset made for the reaseach presented in my paper about [\"Person-Independent sEMG Gesture Recognition Using LSTM Networks for Human-Computer Interaction\"](https://wiki.tum.de/display/infar/%5B19WS+-+GR%5D+Person-Independent+sEMG+Gesture+Recognition+Using+LSTM+Networks+for+Human-Computer+Interaction).\n\nThis repository contains sEMG Data of 13 subjects recorded with the Myo Armband. Subjects were instructed to wear the armband on their dominant arm, like this:\n\n![image](.github/images/myo-position.png)\n\nThe gestures that were recorded, are the ones from the Rock, Paper, Scissors game:\n\n![Rock](.github/images/rock.png)\n![Paper](.github/images/paper.png)\n![Scissors](.github/images/scissors.png)\n\n## File Naming Convention\n\nEach recording was given it's own folder.\nThe folder are named like this: `\u003csubjectId\u003e_\u003carm\u003e_\u003cindex\u003e` (also called the label)\nWhere arm `r` stands for the right arm.\n\n## Recorded Data\n\nEach kind of data is stored in it's own file per recording.\n| File Ending | Description |\n| - | - |\n|`\u003clabel\u003e-\u003cgestureType\u003e-\u003cindex\u003e-emg.csv`|Raw EMG Data from the Myo armband of the eight sensors|\n|`\u003clabel\u003e-\u003cgestureType\u003e-\u003cindex\u003e-pose.csv`|Pose data (detected poses by the built-in Myo detecion)|`s6_r_1-paper-4-pose.csv`|\n|`\u003clabel\u003e-\u003cgestureType\u003e-\u003cindex\u003e-orientation.csv`|IMU Orientation data|\n|`\u003clabel\u003e-\u003cgestureType\u003e-\u003cindex\u003e-meta.csv`|Meta Data (see below)|\n\n## Meta Data\n\nEach gesture recording comes with a meta data description file.\n\nExample with explainatory comments:\n```yaml\nlabel: s6_r_1 # same as the folder\ngesture: scissors # which gesture the user was told to record\nindex: 8 # 8th gesture of type scissors from this session\n\ndate: 21/08/2020\ntime: 05:49:30\n\narm: 0 # arm 0 = right, arm 1 = left\narm_direction: 1 # -1 if the armband was worn in reverse\nrssi: -53 # signal strength\nbattery_level: 97 # 100 means full battery\nmirror_left_arm: True # wether the recording applicatin automatically mirror the data if the user was left handed\n\nfiles: # which data was recorded; e.g. pose data might miss, if no pose was detected\n  - s6_r_1-scissors-8-pose.csv\n  - s6_r_1-scissors-8-orientation.csv\n  - s6_r_1-scissors-8-emg.csv\n\n```\n\n## Data Recording Procedure\n\nIn order to generate labeled data, a Python application\ninterfacing with the Myo was implemented. The data is recorded while\nshowing an icon and text label describing the gesture. Each gesture is\nrecorded for two seconds following a 0.5 second break where subjects\nare supposed to relax their arm. The two-second time-span was chosen\nbecause several tests showed that it is the amount of time required for\nmost persons to perform the transition from rest pose to the wanted\ngesture and back. In one recording iteration, each gesture is recorded\nten times, which makes a total of 30 gestures per iteration. Therefore,\none iteration takes 75 seconds. Those values were chosen after initial\ntesting, which showed that subjects quickly get exhausted performing the\ngestures. After a short break, without taking the armband off, a second\niteration was performed in the same session. This is then repeated once\nmore. Some participants were recorded a second time: After a long break,\nthe armband was put on differently, and the process described above was\nrepeated. In total, each participant contributed three (one session) or\nsix iterations (two sessions). One additional participant recorded 30\ndatasets in 10 sessions on different days to create a highly specialized\ndataset.\n\nThirteen healthy subjects aged between 18 and 58 years participated in\nthe data acquisition process. Four of them were female, and nine\ncandidates male. One participant was left-handed. All subjects agreed to\nthe data recording and analysis by writing. Before starting the\nrecording session, participants filled out the consent form, which\nstates what is being recorded and how the data is used. The Myo armband\nwas then put onto the thickest part of the forearm, just below the\nelbow on the dominant arm. Then the subject waited until the\nconnection icon on the armband stopped blinking, which indicated that\nthe armband warmed up. After performing the Myo sync gesture, which is\nneeded to detect the arm wearing the armband, the Myo is ready to use.\nSubjects were told to start from a rest pose, performing the gesture and\nreaching the highest intensity when the progress bar indicates that half\nthe time passed.\n\nThere are multiple reasons for always using the dominant arm during the\nrecording. First, a possible user interface using one Myo armband would\nbe controlled with the dominant arm because it feels more natural to the\nuser. But also not using the dominant arm could introduce irregularities\ninto the dataset because users are not used to performing similar\ngestures with their non-dominant arm. Also, differences in the data\nresulting from recording the left and right arm are non-existent since\nthe current arm is detected, and the data mirrored accordingly.\n\nAll data provided by the Myo armband is stored in multiple files labeled\nwith a unique but randomized identifier. Each file contains roughly\n400 ± 5 entries since the recording took two seconds, and the data was\nsampled with. Each entry contains an index, a timestamp, and one value\nfor each of the eight sensors. and other data along with\nmeta-information (label, recording time, left or right arm, battery\nlevel, and connection quality) is stored in other files.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmichidk%2Fmyo-dataset","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fmichidk%2Fmyo-dataset","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmichidk%2Fmyo-dataset/lists"}