Limitations of (Procrustes) Alignment in Assessing Multi-Person Human Pose and Shape Estimation
This addresses the challenge of evaluating multi-person human pose and shape estimation in surveillance, but appears incremental as it builds on existing metrics and alignment techniques.
The paper tackles the problem of accurately estimating 3D human pose and shape in video surveillance by advocating for metrics like W-MPJPE and W-PVE that omit Procrustes alignment and introducing RotAvat to refine mesh alignment with the ground plane, demonstrating its effectiveness through qualitative comparisons.
We delve into the challenges of accurately estimating 3D human pose and shape in video surveillance scenarios. Beginning with the advocacy for metrics like W-MPJPE and W-PVE, which omit the (Procrustes) realignment step, to improve model evaluation, we then introduce RotAvat. This technique aims to enhance these metrics by refining the alignment of 3D meshes with the ground plane. Through qualitative comparisons, we demonstrate RotAvat's effectiveness in addressing the limitations of existing aproaches.