Research map / Detection and segmentation / Open-set and few-shot detection
Objaverse: A Universe of Annotated 3D Objects
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 |
|---|---|---|---|
| Add Objaverse 3D copy-paste augmentation to GOL | Mask-RCNN ResNet-50 on LVIS · LVIS | AP: 27.7 -> 28.3 | Helped |
| Add Objaverse 3D copy-paste augmentation to GOL | Mask-RCNN ResNet-50 on LVIS · LVIS rare categories | APr: 21.4 -> 21.8 | Helped |
| Add Objaverse 3D copy-paste augmentation to GOL | Mask-RCNN ResNet-50 on LVIS · LVIS common categories | APc: 27.7 -> 28.3 | Helped |
| Add Objaverse 3D copy-paste augmentation to GOL | Mask-RCNN ResNet-50 on LVIS · LVIS frequent categories | APf: 30.4 -> 31.1 | Helped |
| Add Objaverse 3D copy-paste augmentation to GOL | Mask-RCNN ResNet-50 on LVIS detection · LVIS detection | AP: 27.5 -> 28.9 | Helped |
| Add Objaverse 3D copy-paste augmentation to GOL | Mask-RCNN ResNet-50 on LVIS detection · LVIS rare-category detection | APr: 19.8 -> 21.8 | Helped |
| Train GET3D on Objaverse bags rather than ShapeNet bags | GET3D · Bag generation diversity | human preference for greater diversity: 91% | Helped |
| Increase ObjectNav training from 18 million to approximately 460 million simulation steps | Objaverse-trained ObjectNav agent · Open-Vocabulary ObjectNav | success rate: 19.9% -> 33.0% | Helped |
| Use an Objaverse-trained ObjectNav agent instead of a random policy | Open-Vocabulary ObjectNav · Open-Vocabulary ObjectNav | success rate: 5.1% -> 19.9% | Helped |
| Evaluate CLIP on random object orientations rather than the best-performing orientation | LAION-2B ViT-L/14 · Rotational robustness benchmark | Top-1 accuracy: 52.1% -> 32.9% | Hurt |
Closest papers by meaning
- ImageNet3D: Towards General-Purpose Object-Level 3D Understanding (NeurIPS 2024 · similarity 0.88)
- Towards 3D Objectness Learning in an Open World (NeurIPS 2025 · similarity 0.88)
- One-shot 3D Object Canonicalization based on Geometric and Semantic Consistency (CVPR 2025 · similarity 0.88)
- OpenLex3D: A Tiered Benchmark for Open-Vocabulary 3D Scene Representations (NeurIPS 2025 · similarity 0.88)
- ObjVariantEnsemble: Advancing Point Cloud LLM Evaluation in Challenging Scenes with Subtly Distinguished Objects (AAAI 2025 · similarity 0.87)
- Leveraging VLM-Based Pipelines to Annotate 3D Objects (ICML 2024 · 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.