Search papers, labs, and topics across Lattice.
This paper introduces a monocular video pipeline leveraging DreamPose3D with a drift-controlled global lifting module to recover 18 clinically relevant biomechanics metrics from broadcast baseball footage. The system incorporates a kinematics refinement pipeline to address motion blur and compression artifacts, ensuring temporally stable and physically plausible kinematics. Validated on 156 pitches from 13 professional pitchers, the pipeline achieves sub-degree agreement with gold-standard measurements and demonstrates AUCs of 0.811 and 0.825 for predicting Tommy John surgery and significant arm injuries, respectively, on a larger dataset of 7,348 pitchers.
Unlock scalable injury-risk screening in baseball with a monocular video pipeline that rivals expensive stadium-scale motion capture, achieving high accuracy in predicting arm injuries from broadcast footage.
Injury prediction in pitching depends on precise biomechanical signals, yet gold-standard measurements come from expensive, stadium-installed multi-camera systems that are unavailable outside professional venues. We present a monocular video pipeline that recovers 18 clinically relevant biomechanics metrics from broadcast footage, positioning pose-derived kinematics as a scalable source for injury-risk modeling. Built on DreamPose3D, our approach introduces a drift-controlled global lifting module that recovers pelvis trajectory via velocity-based parameterization and sliding-window inference, lifting pelvis-rooted poses into global space. To address motion blur, compression artifacts, and extreme pitching poses, we incorporate a kinematics refinement pipeline with bone-length constraints, joint-limited inverse kinematics, smoothing, and symmetry constraints to ensure temporally stable and physically plausible kinematics. On 13 professional pitchers (156 paired pitches), 16/18 metrics achieve sub-degree agreement (MAE $<1^{\circ}$). Using these metrics for injury prediction, an automated screening model achieves AUC 0.811 for Tommy John surgery and 0.825 for significant arm injuries on 7,348 pitchers. The resulting pose-derived metrics support scalable injury-risk screening, establishing monocular broadcast video as a viable alternative to stadium-scale motion capture for biomechanics.
