# Top-down heatmap-based pose estimation Top-down methods divide the task into two stages: object detection, followed by single-object pose estimation given object bounding boxes Instead of estimating keypoint coordinates directly, the pose estimator will produce heatmaps which represent the likelihood of being a keypoint, following the paradigm introduced in [Simple Baselines for Human Pose Estimation and Tracking](http://openaccess.thecvf.com/content_ECCV_2018/html/Bin_Xiao_Simple_Baselines_for_ECCV_2018_paper.html).
## Results and Models ### Animal-Pose Dataset Results on AnimalPose validation set (1117 instances) | Model | Input Size | AP | AR | Details and Download | | :--------: | :--------: | :---: | :---: | :-------------------------------------------------------: | | HRNet-w32 | 256x256 | 0.740 | 0.780 | [hrnet_animalpose.md](./animalpose/hrnet_animalpose.md) | | HRNet-w48 | 256x256 | 0.738 | 0.778 | [hrnet_animalpose.md](./animalpose/hrnet_animalpose.md) | | ResNet-152 | 256x256 | 0.704 | 0.748 | [resnet_animalpose.md](./animalpose/resnet_animalpose.md) | | ResNet-101 | 256x256 | 0.696 | 0.736 | [resnet_animalpose.md](./animalpose/resnet_animalpose.md) | | ResNet-50 | 256x256 | 0.691 | 0.736 | [resnet_animalpose.md](./animalpose/resnet_animalpose.md) | ### AP-10K Dataset Results on AP-10K validation set | Model | Input Size | AP | Details and Download | | :--------: | :--------: | :---: | :--------------------------------------------------: | | HRNet-w48 | 256x256 | 0.728 | [hrnet_ap10k.md](./ap10k/hrnet_ap10k.md) | | HRNet-w32 | 256x256 | 0.722 | [hrnet_ap10k.md](./ap10k/hrnet_ap10k.md) | | ResNet-101 | 256x256 | 0.681 | [resnet_ap10k.md](./ap10k/resnet_ap10k.md) | | ResNet-50 | 256x256 | 0.680 | [resnet_ap10k.md](./ap10k/resnet_ap10k.md) | | CSPNeXt-m | 256x256 | 0.703 | [cspnext_udp_ap10k.md](./ap10k/cspnext_udp_ap10k.md) | ### Desert Locust Dataset Results on Desert Locust test set | Model | Input Size | AUC | EPE | Details and Download | | :--------: | :--------: | :---: | :--: | :-------------------------------------------: | | ResNet-152 | 160x160 | 0.925 | 1.49 | [resnet_locust.md](./locust/resnet_locust.md) | | ResNet-101 | 160x160 | 0.907 | 2.03 | [resnet_locust.md](./locust/resnet_locust.md) | | ResNet-50 | 160x160 | 0.900 | 2.27 | [resnet_locust.md](./locust/resnet_locust.md) | ### Grévy’s Zebra Dataset Results on Grévy’s Zebra test set | Model | Input Size | AUC | EPE | Details and Download | | :--------: | :--------: | :---: | :--: | :----------------------------------------: | | ResNet-152 | 160x160 | 0.921 | 1.67 | [resnet_zebra.md](./zebra/resnet_zebra.md) | | ResNet-101 | 160x160 | 0.915 | 1.83 | [resnet_zebra.md](./zebra/resnet_zebra.md) | | ResNet-50 | 160x160 | 0.914 | 1.87 | [resnet_zebra.md](./zebra/resnet_zebra.md) | ### Animal-Kingdom Dataset Results on AnimalKingdom test set | Model | Input Size | class | PCK(0.05) | Details and Download | | :-------: | :--------: | :-----------: | :-------: | :---------------------------------------------------: | | HRNet-w32 | 256x256 | P1 | 0.6323 | [hrnet_animalkingdom.md](./ak/hrnet_animalkingdom.md) | | HRNet-w32 | 256x256 | P2 | 0.3741 | [hrnet_animalkingdom.md](./ak/hrnet_animalkingdom.md) | | HRNet-w32 | 256x256 | P3_mammals | 0.571 | [hrnet_animalkingdom.md](./ak/hrnet_animalkingdom.md) | | HRNet-w32 | 256x256 | P3_amphibians | 0.5358 | [hrnet_animalkingdom.md](./ak/hrnet_animalkingdom.md) | | HRNet-w32 | 256x256 | P3_reptiles | 0.51 | [hrnet_animalkingdom.md](./ak/hrnet_animalkingdom.md) | | HRNet-w32 | 256x256 | P3_birds | 0.7671 | [hrnet_animalkingdom.md](./ak/hrnet_animalkingdom.md) | | HRNet-w32 | 256x256 | P3_fishes | 0.6406 | [hrnet_animalkingdom.md](./ak/hrnet_animalkingdom.md) |