IVT: An End-to-End Instance-guided Video Transformer for 3D Pose Estimation

Benchmark Model Rank Results
3d-human-pose-estimation-on-3dpwIVT (f=5)PA-MPJPE: 46
3d-human-pose-estimation-on-human36mIVT (f=5)Average MPJPE (mm): 40.2Using 2D ground-truth joints: No
3d-multi-person-pose-estimation-on-cmuIVT (f=5)Average MPJPE (mm): 48.4