Research map / 3D vision / 3D generation
VoxDet: Voxel Learning for Novel Instance Detection
Bowen Li, Jiashun Wang, Yaoyu Hu, Chen Wang, Sebastian Scherer
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 |
|---|---|---|---|
| Remove reconstruction pre-training (ablation) | VoxDet · RoboTools | mAR: 18.7 -> 14.2 | Hurt |
| Replace voxel relation with 3D depth-wise convolution matching (ablation) | VoxDet · RoboTools | mAR: 18.7 -> 15.1 | Hurt |
| Remove rotation measurement module R (ablation) | VoxDet · RoboTools | mAR: 18.7 -> 15.6 | Hurt |
| Do not supervise the predicted rotation (ablation) | VoxDet · RoboTools | mAR: 18.7 -> 18.2 | Hurt |
| Voxel matching instead of correlation-based 2D matching | VoxDet · LM-O/YCB-V average | mAR: 23.6 -> 30.4 | Improved |
| Voxel matching instead of DINO cosine similarity | VoxDet · LM-O/YCB-V average | mAR: 24.6 -> 30.4 | Improved |
| Voxel matching instead of correlation | VoxDet · RoboTools (sim-to-real) | mAR: 14.4 -> 18.7 | Improved |
| Voxel matching instead of correlation matching | VoxDet · per-image inference | total time (s): 0.182 -> 0.154 | Improved |
| Voxel matching instead of DINO features | VoxDet · per-image inference | total time (s): 0.357 -> 0.154 | Improved |
| Reduce templates from 10 to 2 | OLN_DINO 2D baseline vs VoxDet · YCB-V | accuracy drop: 87% | Hurt |
| Reduce Top-K from 100 to 10 | VoxDet vs OS2D · Top-K analysis | AR: 5~10% vs 38% | 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.
- Gen6D — Gen6D: Generalizable Model-Free 6-DoF Object Pose Estimation from RGB Images (ECCV 2022) · better on 2 of 2 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
- Voxel R-CNN: Towards High Performance Voxel-based 3D Object Detection (AAAI 2021 · similarity 0.89)
- Voxel-based 3D Detection and Reconstruction of Multiple Objects from a Single Image (NeurIPS 2021 · similarity 0.89)
- VoxDet: Rethinking 3D Semantic Scene Completion as Dense Object Detection (NeurIPS 2025 · similarity 0.89)
- Voxel Mamba: Group-Free State Space Models for Point Cloud based 3D Object Detection (NeurIPS 2024 · similarity 0.88)
- Solving Instance Detection from an Open-World Perspective (CVPR 2025 · similarity 0.86)
- DVMNet: Computing Relative Pose for Unseen Objects Beyond Hypotheses (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.
