Disclosure: I’m one of the people behind Aidlab.<\/p>\n
If you’re working on human activity recognition, exercise classification, or repetition counting, we’ve published AIDLAB-HAR on Hugging Face under CC BY 4.0.<\/p>\n
It contains 180 chest-worn recordings and 518,300 samples at 50 Hz. The signals include 3-axis acceleration, orientation quaternions, 16 activity labels, and repetition markers.<\/p>\n
We packaged it into three configurations that work directly with the One detail we found while preparing the release: the paper’s overview mentions 15 activities, while the archive contains 16 filename labels. We documented the discrepancy and preserved the source labels.<\/p>\n Dataset: https:\/\/huggingface.co\/datasets\/aidlab-wearables\/AIDLAB-HAR<\/a><\/p>\n Paper: https:\/\/doi.org\/10.3390\/s24123891<\/a><\/p>\n<\/div>\n submitted by \/u\/Guzikk <\/a> datasets<\/code> library: recordings<\/code>, signals<\/code>, and annotations<\/code>. The raw v3 archive, cleaning notes, and preparation script are included too.<\/p>\n
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
