Closed felixdivo closed 1 year ago
We do not have a ready-to-use JSON version of this data, but it should be quick for you to create.
Basically, what you're looking for is a mapping between raw_label
and act_cat
in a BABEL annotation (e.g., below). You can loop over each annotation in the dataset, and create a dictionary of {<act_cat>: [ <raw_label_1>, <raw_label_2>, ... ], ... }
"raw_label": "walk back and forth",
"proc_label": "walk back and forth",
"seg_id": "3ca74412-f765-445b-b51b-6c2e72316010",
"act_cat": [
"walk",
"forward movement",
"backwards movement"
]
Hope this answers your question!
Okay, thank you!
@carjun Would I be allowed to publish this extracted data (and nothing else of the data, of course)? E.g. in a MIT-licensed GitHub repo?
Is the "Table of Action categories" available in a simple machine-readable format (JSON, ...) too?