Occlusion-Aware Networks for 3D Human Pose Estimation in Video

Benchmark Model Rank Results
3d-human-pose-estimation-on-human36mOcclusion-Aware NetworksAverage MPJPE (mm): 42.9Using 2D ground-truth joints: No
3d-human-pose-estimation-on-humaneva-iOcclusion-Aware NetworksMean Reconstruction Error (mm): 14.3