Sign in to view source links and access this dataset
Description
A dataset for action spotting in artistic gymnastics, formatted similarly to OpenSportsLab and SoccerNet datasets. It contains video clips and corresponding JSON annotations for events on vault, uneven bars, balance beam, and floor exercise. The dataset was created by OpenSportsLab and was last updated on March 3, 2026.
Use Cases
Training action detection models based on clip-relative annotations for gymnastics events.
Benchmarking temporal localization algorithms on the sport domain of artistic gymnastics.
Developing multi-event classifiers for gymnastics apparatus like vault, uneven bars, balance beam, and floor exercise.
Strengths
Annotations are provided in a structured JSON format for machine readability.
Dataset is organized into standard splits (train, valid, test) for model development.
Focuses on a specific sport domain (artistic gymnastics) with four distinct apparatus.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
OpenSportsLab
Freshness
Last updated 2026-03-03 22:14:43; freshness should be verified.
License is unknown; terms of use must be verified before application.