Label Halpe-Body Landmarks avatar

Label Halpe-Body Landmarks

1 version
Open in App

Convert landmark list to Halpe 26-keypoint full body schema

Use This When

  • Standardizing pose landmarks to Halpe-26 format with foot detail keypoints
  • Building pose analysis requiring ankle, heel, and toe landmarks for gait analysis
  • Interfacing with components expecting Halpe schema including head, neck, hip keypoints
  • Preparing full-body pose data for sports analytics or biomechanics applications

What It Does

  • Converts ordered 26-point landmark list to structured Human26 format with named fields
  • Maps full body including head, neck, hip, and detailed foot keypoints (toes, heels)
  • Validates minimum landmark count and fails if fewer than 26 landmarks provided
  • Returns typed Landmarks2d.Human26 structure for downstream full-body analysis

Works Best With

  • Halpe pose detector → this component → gait analysis or sports biomechanics
  • Full-body pose estimation → this component → pose-based action recognition
  • Integration with visualize for complete skeleton rendering including feet
  • Feeding analytics requiring precise foot placement or balance analysis

Caveats

  • Requires exactly 26 or more landmarks; fails with fewer even if partial pose sufficient
  • No validation that landmark semantics match Halpe ordering; assumes correct input sequence
  • Foot keypoints require high resolution and clear view; occlusion reduces quality
  • Additional landmarks beyond 26 ignored; uses only first 26 points in order

Versions

  • 7a9bdd7elatestdefaultlinux/amd64

    Automated release