Published August 26, 2025 | Version 1.0
Dataset Open

EHG-7: Event-based Hand Gestures-7

  • 1. ROR icon Lodz University of Technology

Description

Dataset description

This dataset contains multi-modal recordings of seven distinct hand gestures for machine learning research. Each gesture was recorded in seven different lighting conditions, with each recording session containing 20 repetitions of the gesture. The data was captured simultaneously using a conventional RGB webcam and an iniVation DVXplorer Mini event camera, making it suitable for developing and benchmarking event-based and multi-modal deep learning models for gesture recognition.

 

Reassemble and Extract

Verify integrity:
sha256sum -c EHG-7.tar.gz.sha256

Reassemble and extract:
cat EHG-7.tar.gz.* | tar -xvzf -

Files

Files (14.9 GB)

Name Size Download all
md5:433fee7eed7783adcff5aca8e0ca3990
4.3 GB Download
md5:cdea1b14fc8cca8779ac0f7f2dc5a61d
4.3 GB Download
md5:91baf0a36ac1bfd48747bc0faac916fb
4.3 GB Download
md5:70f177165764a0d82365c5b919653238
2.0 GB Download
md5:55337650c82325a2ca463e12a65dbfb2
332 Bytes Download

Additional details

Funding

European Commission
MISEL - MULTISPECTRAL INTELLIGENT VISION SYSTEM WITH EMBEDDED LOW-POWER NEURAL COMPUTING 101016734