Generating Multiple Hypotheses for 3D Human Pose Estimation with Mixture Density Network

Benchmark Model Rank Results
3d-human-pose-estimation-on-human36mMDN (Multi-View)#41Average MPJPE (mm): 49.6Using 2D ground-truth joints: No
3d-human-pose-estimation-on-human36mMDN#55Average MPJPE (mm): 52.7Using 2D ground-truth joints: No
3d-human-pose-estimation-on-mpi-inf-3dhpMDM#80PCK: 67.9
monocular-3d-human-pose-estimation-on-human3Multimodal Mixture Density Networks#23Average MPJPE (mm): 52.7Use Video Sequence: NoFrames Needed: 1
multi-hypotheses-3d-human-pose-estimation-on-2SMPL-MDN (by 3D Multi-bodies)#4Best-Hypothesis PMPJPE (n = 25): 69.5H36M PMPJPE (n = 25): 42.7
multi-hypotheses-3d-human-pose-estimation-on-human36mMDN#8Average MPJPE (mm): 52.7Average PMPJPE (mm): 42.6