TCPFormer: Learning Temporal Correlation with Implicit Pose Proxy for 3D Human Pose Estimation

Benchmark Model Rank Results
3d-human-pose-estimation-on-mpi-inf-3dhpTCPFormer (T=81)#1MPJPE: 15AUC: 87.7PCK: 99.0
3d-human-pose-estimation-on-mpi-inf-3dhpTCPFormer (T=27)#6MPJPE: 17.8AUC: 86.5PCK: 98.7
monocular-3d-human-pose-estimation-on-human3-6mTCPFormer (SH, T=243, seq2seq)#5Average MPJPE (mm): 37.9
3d-human-pose-estimation-on-human36mTCPFormer (SH, T=243, seq2seq)#78PA-MPJPE: 31.7