Researchers propose BioModule, a lightweight plug-in temporal transformer that attaches downstream of any 3D pose estimator to predict biomechanical attributes from standard 17-joint skeletons. The model is estimator-agnostic and requires no modification of the upstream pose model.
- BioModule predicts biomechanical quantities using a large-scale aligned dataset pairing Human3.6M video and keypoints with the biomechanical label space of Human3.6Mplus.
- The authors establish anatomical correspondence between coordinate systems for frame-accurate cross-modal supervision.
- They benchmark BioModule across seven state-of-the-art 3D pose estimators to analyze how upstream quality propagates to downstream fidelity.
The work positions BioModule as a compact, modular bridge between vision-based pose estimation and biomechanically meaningful human motion analysis.