Research map / 3D vision / Human pose estimation
DensePose: Dense Human Pose Estimation in the Wild
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use ResNet-101 instead of ResNet-50 (ablation) | DensePose-RCNN · COCO minival | AP: 51.0 -> 51.8 | Helped |
| Add a mask branch (ablation) | DensePose-RCNN · COCO minival | AP: 51.0 -> 51.9 | Helped |
| Add a keypoint branch (ablation) | DensePose-RCNN · COCO minival | AP: 51.0 -> 52.8 | Helped |
| Add iterative refinement without auxiliary-task input (ablation) | DensePose-RCNN · COCO minival | AP: 51.0 -> 51.6 | Helped |
| Add cascading to the mask-branch model (ablation) | DensePose + masks · COCO minival | AP: 51.9 -> 52.8 | Helped |
| Add cascading to the keypoint-branch model (ablation) | DensePose + keypoints · COCO minival | AP: 52.8 -> 55.8 | Helped |
| Train with dense inpainted supervision instead of sparse manually annotated points (ablation) | DensePose-RCNN · COCO multi-person dense correspondence | RCP AUC | Helped |
| Use region-based processing instead of a fully-convolutional network (ablation) | DensePose · COCO multi-person dense correspondence | RCP AUC | Helped |
| Train on DensePose rather than surrogate datasets (ablation) | ResNet-101 FCN · Single-person dense pose estimation | RCP AUC | Helped |
Closest papers by meaning
- DenseReg: Fully Convolutional Dense Shape Regression In-The-Wild (CVPR 2017 · similarity 0.91)
- Dense Human Body Correspondences Using Convolutional Networks (CVPR 2016 · similarity 0.91)
- BodyMap: Learning Full-Body Dense Correspondence Map (CVPR 2022 · similarity 0.90)
- MeshPose: Unifying DensePose and 3D Body Mesh Reconstruction (CVPR 2024 · similarity 0.88)
- MultiPly: Reconstruction of Multiple People from Monocular Video in the Wild (CVPR 2024 · similarity 0.87)
- ODHSR: Online Dense 3D Reconstruction of Humans and Scenes from Monocular Videos (CVPR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.