Human Pose Estimation on COCO keypoint (val)
75.5APHR-NAS-B
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| HR-NAS-BInput size=384x288, Params=6.6M, FLOPs=3.72G2021.06 | 75.5 | 72.6 | 81.7 | 79.4 | — | — | |
| HRNet-W32Input size=384x288, Params=28.5M, FLOPs=16.0G2021.06 | 74.9 | 71.5 | 80.8 | 79.3 | — | — | |
| HR-NAS-BInput size=256x192, Params=6.1M, FLOPs=1.35G2021.06 | 73.7 | 70.2 | 80.6 | 79.3 | — | — | |
| AutoPoseInput size=256x192, FLOPs=10.65G2021.06 | 73.6 | 69.8 | 79.7 | 78.1 | — | — | |
| HRNet-W32Input size=256x192, Params=28.5M, FLOPs=7.10G2021.06 | 73.4 | 70.2 | 80.1 | 78.9 | — | — | |
| PoseNFS-3Input size=384x288, Params=15.8M, FLOPs=14.8G2021.06 | 73 | — | — | — | — | — | |
| SimpleBaseline-50Input size=384x288, Params=34.0M, FLOPs=20.02G, ImageNet Pre-trained=true2021.06 | 72.2 | 68.1 | 79.7 | 77.6 | — | — | |
| SimpleBaseline-50Input size=256x192, Params=34.0M, FLOPs=8.90G, ImageNet Pre-trained=true2021.06 | 70.4 | 67.1 | 77.2 | 76.3 | — | — | |
| CPNInput size=256x192, Params=27.0M, FLOPs=6.20G2021.06 | 69.4 | — | — | — | — | — | |
| DA-NASInput size=256x192, Params=10.9M, FLOPs=2.18G2021.06 | 68.4 | 65.5 | 74.4 | 75.7 | — | — | |
| RTMPoseMethod Category=Task-specific Specialist, Evaluation Paradigm=Top-down2024.03 | 68.2 | — | — | — | 88.3 | 75.9 | |
| HR-NAS-AInput size=256x192, Params=1.7M, FLOPs=0.25G2021.06 | 67.7 | 65.4 | 71.1 | 70.8 | — | — | |
| MobileNetV2Input size=384x288, Params=2.3M, FLOPs=0.74G, ImageNet Pre-trained=true2021.06 | 67.3 | 62.8 | 74.7 | 72.8 | — | — | |
| Lumen-7BMethod Category=LMM Generalist, Evaluation Paradigm=Top-down2024.03 | 67.2 | — | — | — | 90.4 | 75.6 | |
| NAS-CSSInput size=256x192, Params=2.9M, FLOPs=1.48G2021.06 | 65.9 | 63.1 | 70 | 69.3 | — | — | |
| HR-NAS-AInput size=384x288, Params=1.1M, FLOPs=0.35G2021.06 | 65.7 | 62.5 | 72.1 | 71.4 | — | — | |
| Pix2Seq-v2Method Category=Vision Generalist, Evaluation Paradigm=Top-down2024.03 | 64.8 | — | — | — | — | — | |
| MobileNetV2Input size=256x192, Params=2.3M, FLOPs=0.33G, ImageNet Pre-trained=true2021.06 | 64.6 | 61 | 71.1 | 70.7 | — | — | |
| ShuffleNetV2Input size=384x288, Params=1.3M, FLOPs=0.37G, ImageNet Pre-trained=true2021.06 | 63.6 | 59.5 | 70.7 | 69.7 | — | — | |
| CPMMethod Category=Task-specific Specialist, Evaluation Paradigm=Top-down2024.03 | 62.7 | — | — | — | 86.2 | 70.9 | |
| ShuffleNetV1Input size=384x288, Params=1.0M, FLOPs=0.35G, ImageNet Pre-trained=true2021.06 | 62.2 | 57.8 | 69.5 | 68.4 | — | — | |
| ShuffleNetV2Input size=256x192, Params=1.3M, FLOPs=0.17G, ImageNet Pre-trained=true2021.06 | 59.8 | 56.5 | 66.2 | 66.4 | — | — | |
| ShuffleNetV1Input size=256x192, Params=1.0M, FLOPs=0.16G, ImageNet Pre-trained=true2021.06 | 58.5 | 55.2 | 64.6 | 65.1 | — | — |