Tasks SotA Datasets Papers Methods Submit About

Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable Benchmarks All SotA Datasets Papers Methods

Community

Submit Results About

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/MIPNet (HRNet-W48)

MIPNet (HRNet-W48)

Reported on 30 benchmarks across 5 tasks · 1 paper · 11 SOTA

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

Computer Vision15 results

Pose EstimationonOCHuman
Test AP· 2021-01-27
42.5
best: 93.3 (ViTPose (ViTAE-G, GT bounding boxes))
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Pose EstimationonCrowdPose
AP· 2021-01-27
70
best: 78.5 (BUCTD-W48 (w/cond. input from PETR, and generative sampling))
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Pose EstimationonCrowdPose
APM· 2021-01-27
71.1
best: 86.6 (ViTPose-G)
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Pose EstimationonOCHuman
Validation AP· 2021-01-27
42
best: 92.8 (ViTPose (ViTAE-G, GT bounding boxes))
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Pose EstimationonCrowdPose
AP Hard· 2021-01-27
59.4
best: 466 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Pose EstimationonOCHuman
Test AP· 2021-01-27
42.5
best: 93.3 (ViTPose (ViTAE-G, GT bounding boxes))
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Pose EstimationonOCHuman
Validation AP· 2021-01-27
42
best: 92.8 (ViTPose (ViTAE-G, GT bounding boxes))
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Pose EstimationonCrowdPose
AP Easy· 2021-01-27
78.1
best: 88.8 (RTMO-l)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Pose EstimationonCrowdPose
AP Hard· 2021-01-27
59.4
best: 466 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Pose EstimationonCrowdPose
AP Medium· 2021-01-27
71.1
best: 566 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Pose EstimationonCrowdPose
mAP @0.5:0.95· 2021-01-27
70
best: 83.8 (RTMO-l)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Multi-Person Pose EstimationonCrowdPose
AP Easy· 2021-01-27
78.1
best: 88.8 (RTMO-l)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Multi-Person Pose EstimationonCrowdPose
AP Hard· 2021-01-27
59.4
best: 466 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Multi-Person Pose EstimationonCrowdPose
AP Medium· 2021-01-27
71.1
best: 566 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
Multi-Person Pose EstimationonCrowdPose
mAP @0.5:0.95· 2021-01-27
70
best: 83.8 (RTMO-l)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223

Methodology11 results

3DonOCHuman
Test AP· 2021-01-27
42.5
best: 93.3 (ViTPose (ViTAE-G, GT bounding boxes))
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
3DonCrowdPose
AP· 2021-01-27
70
best: 78.5 (BUCTD-W48 (w/cond. input from PETR, and generative sampling))
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
3DonCrowdPose
APM· 2021-01-27
71.1
best: 86.6 (ViTPose-G)
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
3DonOCHuman
Validation AP· 2021-01-27
42
best: 92.8 (ViTPose (ViTAE-G, GT bounding boxes))
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
3DonCrowdPose
AP Hard· 2021-01-27
59.4
best: 466 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
3DonOCHuman
Test AP· 2021-01-27
42.5
best: 93.3 (ViTPose (ViTAE-G, GT bounding boxes))
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
3DonOCHuman
Validation AP· 2021-01-27
42
best: 92.8 (ViTPose (ViTAE-G, GT bounding boxes))
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
3DonCrowdPose
AP Easy· 2021-01-27
78.1
best: 88.8 (RTMO-l)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
3DonCrowdPose
AP Hard· 2021-01-27
59.4
best: 466 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
3DonCrowdPose
AP Medium· 2021-01-27
71.1
best: 566 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
3DonCrowdPose
mAP @0.5:0.95· 2021-01-27
70
best: 83.8 (RTMO-l)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223

Audio11 results

1 Image, 2*2 StitchionOCHuman
Test AP· 2021-01-27
42.5
best: 93.3 (ViTPose (ViTAE-G, GT bounding boxes))
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
1 Image, 2*2 StitchionCrowdPose
AP· 2021-01-27
70
best: 78.5 (BUCTD-W48 (w/cond. input from PETR, and generative sampling))
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
1 Image, 2*2 StitchionCrowdPose
APM· 2021-01-27
71.1
best: 86.6 (ViTPose-G)
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
1 Image, 2*2 StitchionOCHuman
Validation AP· 2021-01-27
42
best: 92.8 (ViTPose (ViTAE-G, GT bounding boxes))
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
1 Image, 2*2 StitchionCrowdPose
AP Hard· 2021-01-27
59.4
best: 466 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
1 Image, 2*2 StitchionOCHuman
Test AP· 2021-01-27
42.5
best: 93.3 (ViTPose (ViTAE-G, GT bounding boxes))
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
1 Image, 2*2 StitchionOCHuman
Validation AP· 2021-01-27
42
best: 92.8 (ViTPose (ViTAE-G, GT bounding boxes))
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
1 Image, 2*2 StitchionCrowdPose
AP Easy· 2021-01-27
78.1
best: 88.8 (RTMO-l)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
1 Image, 2*2 StitchionCrowdPose
AP Hard· 2021-01-27
59.4
best: 466 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
1 Image, 2*2 StitchionCrowdPose
AP Medium· 2021-01-27
71.1
best: 566 (DETRPose-N)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
1 Image, 2*2 StitchionCrowdPose
mAP @0.5:0.95· 2021-01-27
70
best: 83.8 (RTMO-l)
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223

Knowledge Base2 results

2D Human Pose EstimationonOCHuman
Test AP· 2021-01-27
42.5
best: 48.3 (BBox-Mask-Pose 2x)
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223
2D Human Pose EstimationonOCHuman
Validation AP· 2021-01-27
42
best: 48.6 (BBox-Mask-Pose 2x)
SOTA
Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation arXiv:2101.11223