3D-POP - An automated annotation approach to facilitate markerless 2D-3D tracking of freely moving birds with marker-based motion capture

The 3DPOP dataset is a large scale 2D to 3D posture, identity and trajectory dataset for freely moving pigeons. We use marker-based motion tracking to first track precise head and body position and orientation for multiple individuals, then propagated custom keypoints based on the relative positions of markers and keypoints.