Skip to content

Question About the Post-Processing Details for Human Demonstration Data #47

Description

@tannq-jo

Hi, I really enjoyed reading your paper. I found it very interesting. I would like to ask about the details of the data processing pipeline for human demonstration data. Could you please explain or elaborate more on the following steps?

  • Motion-Quality Filtering: We ensure data reliability by filtering noisy HaWoR estimates based on detection confidence and DBA error, while removing high-frequency jitter or discontinuities in wrist-space.
  • Handedness Debiasing: We apply left-right spatial mirroring to all samples to mitigate common right-handed bias, thereby promoting ambidextrous generalization in the learned policy.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions