Occlusion-Aware Networks

Reported on 8 benchmarks across 4 tasks

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision4 results

3D Human Pose EstimationonHumanEva-I
Mean Reconstruction Error (mm)
14.3
best: 9.2 (GLA-GCN (T=27, GT))
3D Human Pose EstimationonHuman3.6M
Average MPJPE (mm)
42.9
best: 131.7 (Rhodin et al.)
Pose EstimationonHumanEva-I
Mean Reconstruction Error (mm)
14.3
best: 9.2 (GLA-GCN (T=27, GT))
Pose EstimationonHuman3.6M
Average MPJPE (mm)
42.9
best: 131.7 (Rhodin et al.)

Methodology2 results

3DonHumanEva-I
Mean Reconstruction Error (mm)
14.3
best: 9.2 (GLA-GCN (T=27, GT))
3DonHuman3.6M
Average MPJPE (mm)
42.9
best: 131.7 (Rhodin et al.)

Audio2 results

1 Image, 2*2 StitchionHumanEva-I
Mean Reconstruction Error (mm)
14.3
best: 9.2 (GLA-GCN (T=27, GT))
1 Image, 2*2 StitchionHuman3.6M
Average MPJPE (mm)
42.9
best: 131.7 (Rhodin et al.)