Research map / Label-efficient and robust learning / Domain adaptation and generalization
Unsupervised Domain Adaptation for 3D Keypoint Estimation via View Consistency
Xingyi Zhou, Arjun Karpur, Chuang Gan, Linjie Luo, Qixing Huang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add view-consistency term (ablation) | full method · Redwood depth (chair) | average distance error (%): 13.95 -> 12.76 | Improved |
| add view-consistency term (ablation) | full method · ShapeNet (chair) | average distance error (%): 6.70 -> 6.60 | Improved |
| add geometric alignment term (ablation) | full method · Redwood depth (chair) | average distance error (%): 12.97 -> 12.76 | Improved |
| add geometric alignment term (ablation) | full method · ShapeNet (chair) | average distance error (%): 6.67 -> 6.60 | Improved |
| update latent configurations by alternating minimization instead of re-initializing (ablation) | full method · Redwood depth (chair) | average distance error (%): 13.43 -> 12.76 | Improved |
| update latent configurations by alternating minimization instead of re-initializing (ablation) | full method · ShapeNet (chair) | average distance error (%): 6.66 -> 6.60 | Improved |
| use proposed adaptation instead of ADDA | ResNet50 keypoint predictor · Redwood depth (chair) | average distance error (%): 15.44 -> 12.76 | Improved |
| use proposed adaptation instead of ADDA | ResNet50 keypoint predictor · ShapeNet (chair) | average distance error (%): 6.98 -> 6.60 | Improved |
| apply ADDA instead of no adaptation | ResNet50 keypoint predictor · 3DCNN depth (chair) | average distance error (%): 11.61 -> 11.81 | No clear effect |
| apply proposed domain adaptation | ResNet50 keypoint predictor · Redwood depth (chair) | average distance error (%): 16.01 -> 12.76 | Improved |
| apply proposed domain adaptation | ResNet50 keypoint predictor · Redwood RGB (chair) | average distance error (%): 27.59 -> 25.24 | Improved |
| apply proposed adaptation to motorcycle keypoints | ResNet50 keypoint predictor · Redwood depth (motorcycle) | average distance error (%): 21.55% -> 18.92% | Improved |
Closest papers by meaning
- Unsupervised Learning of 3D Semantic Keypoints with Mutual Reconstruction (ECCV 2022 · similarity 0.88)
- Unsupervised Monocular 3D Keypoint Discovery from Multi-View Diffusion Priors (CVPR 2026 · similarity 0.87)
- Unsupervised 3D Structure Inference from Category-Specific Image Collections (CVPR 2024 · similarity 0.87)
- Key-Grid: Unsupervised 3D Keypoints Detection using Grid Heatmap Features (NeurIPS 2024 · similarity 0.87)
- Source-Free and Image-Only Unsupervised Domain Adaptation for Category Level Object Pose Estimation (ICLR 2024 · similarity 0.86)
- Back to 3D: Few-Shot 3D Keypoint Detection with Back-Projected 2D Features (CVPR 2024 · similarity 0.86)
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.
