Research map / 3D vision / Human pose estimation
Unified Category-Level Object Detection and Pose Estimation from RGB Images using 3D Prototypes
Tom Fischer, Xiaojie Zhang, Eddy Ilg
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 adapters to DINOv2 blocks (ablation) | Ours · REAL275 | NIoU50: -21.0 | Improved |
| add CrossAttention foreground modeling (mask ignored at inference) (ablation) | Ours · REAL275 | NIoU50: -9.4 -> -4.9 | Improved |
| use the foreground mask at inference (t1=0.5) (ablation) | Ours · REAL275 | NIoU50: -4.9 -> 0 | Improved |
| add instance-level pose refinement (ablation) | Ours · REAL275 | NIoU50: -2.7 -> -2.4 | Improved |
| add size optimization from instance correspondences (ablation) | Ours · REAL275 | NIoU75: -3.9 -> +0.1 | Improved |
| use tuned threshold t2 optimum for no-CA model (ablation) | Ours w/o CA · REAL275 | NIoU50: -9.4 -> -4.9 | Improved |
| single-stage instead of two-stage under image corruptions (ablation) | Ours vs LaPose · REAL275 corruptions | NIoU50 loss: -14.0% vs -24.6% | Improved |
| replace Mask-RCNN ROIs with prototype masks (ablation) | Ours · REAL275 | NIoU50: 48.8 -> 53.7 | Improved |
| use ground-truth prototype masks instead of own detection (ablation) | Ours · REAL275 | NIoU50: 53.7 -> 55.3 | Improved |
| train scale network on prototype projections instead of 2D boxes (ablation) | Ours · REAL275 absolute | 5deg 10cm: 17.7 vs 13.6 | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- LaPose — LaPose: Laplacian Mixture Shape Modeling for RGB-Based Category-Level Object Pose Estimation (ECCV 2024) · better on 16 of 17 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- SchemaPose: RGB-Based Category-Level Pose Estimation with Parametric Category Schema (NeurIPS 2026 · similarity 0.91)
- Universal Features Guided Zero-Shot Category-Level Object Pose Estimation (AAAI 2025 · similarity 0.90)
- Unsupervised Learning of Category-Level 3D Pose from Object-Centric Videos (CVPR 2024 · similarity 0.89)
- GIVEPose: Gradual Intra-class Variation Elimination for RGB-based Category-Level Object Pose Estimation (CVPR 2025 · similarity 0.89)
- MFOS: Model-Free & One-Shot Object Pose Estimation (AAAI 2024 · similarity 0.88)
- Co-op: Correspondence-based Novel Object Pose Estimation (CVPR 2025 · similarity 0.88)
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.
