FS-Jump3D Pose Estimator (Tanaka et al.)

Depth-aware pose estimator trained on FS-Jump3D motion capture to recover 3D joint coordinates from monocular broadcast.

Primary tasks
3D pose estimation, Element detection
Modalities
RGB video, 3D pose sequences
Architecture
Hybrid CNN-transformer with depth regression head
Frameworks
PyTorch
Availability
research
Maintainer
Keio University
Released
Mar 2023

Highlights

Implementation Notes

The released checkpoints expect calibrated camera intrinsics; for broadcast feeds, approximate parameters using rink geometry. Pair with sensor data from Tanaka et al. (2023) to validate landing forces.