MuscleMimic GMR Retargeted Motions provides 1,190 pre-processed motion capture sequences for the MyoFullBody musculoskeletal model. The dataset, created by amathislab and last updated in April 2026, groups motions from the KIT locomotion training and test sets and AMASS action transitions. It was generated using General Motion Retargeting (GMR) to align motions with the specific model.
Use Cases
- Training locomotion models based on the KIT training set of 972 motions.
- Testing biomechanical model performance based on the KIT test set of 108 motions.
- Studying action transitions in human movement based on the 110 AMASS transition motions.
- Developing motion retargeting algorithms using the General Motion Retargeting (GMR) method mentioned in the description.
Strengths
- Contains 1,190 total motion sequences across three defined groups.
- Data is pre-retargeted for the MyoFullBody model, potentially reducing preprocessing effort.
- Includes a clear split between training (972 motions) and testing (108 motions) sets for locomotion.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown for individual motion files, which may limit suitability assessment.
- Description metadata is limited; actual data quality and file formats require manual inspection after download.
Provenance
- Source
- amathislab on Hugging Face.
- Collection Method
- Generated using General Motion Retargeting (GMR) from source motion capture data.
- Time Range
- null
- Freshness
- Last updated 2026-04-07 12:19:37; freshness should be verified.
- Geography
- null