title={Microsoft coco: Common objects in context},
author={Lin, Tsung-Yi and Maire, Michael and Belongie, Serge and Hays, James and Perona, Pietro and Ramanan, Deva and Doll{\'a}r, Piotr and Zitnick, C Lawrence},
booktitle={European conference on computer vision},
pages={740--755},
year={2014},
organization={Springer}
}
```
</details>
Results on COCO val2017 without multi-scale test
| Arch | Input Size | AP | AP<sup>50</sup> | AP<sup>75</sup> | AR | AR<sup>50</sup> | ckpt | log |
title={Bottom-up human pose estimation via disentangled keypoint regression},
author={Geng, Zigang and Sun, Ke and Xiao, Bin and Zhang, Zhaoxiang and Wang, Jingdong},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={14676--14686},
year={2021}
}
```
</details>
DEKR is a popular 2D bottom-up pose estimation approach that simultaneously detects all the instances and regresses the offsets from the instance centers to joints.
In order to predict the offsets more accurately, the offsets of different joints are regressed using separated branches with deformable convolutional layers. Thus convolution kernels with different shapes are adopted to extract features for the corresponding joint.
title={Microsoft coco: Common objects in context},
author={Lin, Tsung-Yi and Maire, Michael and Belongie, Serge and Hays, James and Perona, Pietro and Ramanan, Deva and Doll{\'a}r, Piotr and Zitnick, C Lawrence},
booktitle={European conference on computer vision},
pages={740--755},
year={2014},
organization={Springer}
}
```
</details>
Results on COCO val2017 without multi-scale test
| Arch | Input Size | AP | AP<sup>50</sup> | AP<sup>75</sup> | AR | AR<sup>50</sup> | ckpt | log |
title={Microsoft coco: Common objects in context},
author={Lin, Tsung-Yi and Maire, Michael and Belongie, Serge and Hays, James and Perona, Pietro and Ramanan, Deva and Doll{\'a}r, Piotr and Zitnick, C Lawrence},
booktitle={European conference on computer vision},
pages={740--755},
year={2014},
organization={Springer}
}
```
</details>
Results on COCO val2017.
| Arch | BackBone | AP | AP<sup>50</sup> | AP<sup>75</sup> | AR | AR<sup>50</sup> | ckpt | log |
The checkpoint is converted from the official repo. The training of EDPose is not supported yet. It will be supported in the future updates.
The above config follows [Pure Python style](https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta). Please install `mmengine>=0.8.2` to use this config.
Top-down methods divide the task into two stages: object detection, followed by single-object pose estimation given object bounding boxes. At the 2nd stage, integral regression based methods use a simple integral operation relates and unifies the heatmap and joint regression differentiably, thus obtain the keypoint coordinates given the features extracted from the bounding box area, following the paradigm introduced in [Integral Human Pose Regression](https://arxiv.org/abs/1711.08229).
## Results and Models
### COCO Dataset
Results on COCO val2017 with detector having human AP of 56.4 on COCO val2017 dataset
| Model | Input Size | AP | AR | Details and Download |
