Research map / Detection and segmentation / 3D detection and driving
AnnofreeOD: Detecting All Classes at Low Frame Rates Without Human Annotations
Boyi Sun, Yuhang Liu, Houxin He, Yonglin Tian, Fei-Yue Wang
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 |
|---|---|---|---|
| use 2D-to-3D distillation without scene flow instead of UNION | AnnofreeOD · nuScenes val (3-class) | mAP: 25.1 -> 34.4 | Improved |
| use 2D-to-3D distillation without scene flow instead of UNION | AnnofreeOD · nuScenes val (3-class) | NDS: 24.4 -> 36.6 | Improved |
| use AnnofreeOD instead of UNION for class-agnostic detection | AnnofreeOD · nuScenes val (class-agnostic) | AP: 39.5 -> 45.5 | Improved |
| use AnnofreeOD instead of UNION for class-agnostic detection | AnnofreeOD · nuScenes val (class-agnostic) | NDS: 31.7 -> 35.8 | Improved |
| use AnnofreeOD instead of UNION for class-agnostic detection | AnnofreeOD · nuScenes val (class-agnostic) | AOE: 0.876 -> 1.441 | Hurt |
| use AnnofreeOD instead of UNION for class-agnostic detection | AnnofreeOD · nuScenes val (class-agnostic) | AVE: 0.837 -> 1.956 | Hurt |
| use AnnofreeOD instead of UNION for class-agnostic detection | AnnofreeOD · nuScenes val (class-agnostic) | ATE: 0.590 -> 0.328 | Improved |
| compare annotation-free AnnofreeOD with supervised 10-class detector | AnnofreeOD · nuScenes val (10-class) | mAP: 60.0 -> 22.2 | Mixed |
| compare AnnofreeOD with 1% labeled supervised training (3-class) | AnnofreeOD · nuScenes val (3-class) | mAP: 24.3 -> 34.4 | Improved |
| compare AnnofreeOD with 10% labeled supervised training (class-agnostic) | AnnofreeOD · nuScenes val (class-agnostic) | AP: 61.2 -> 45.5 | Hurt |
| compare AnnofreeOD with 10% labeled supervised training (3-class) | AnnofreeOD · nuScenes val (3-class) | mAP: 45.9 -> 34.4 | Hurt |
| use AnnofreeOD instead of UNION for Cyclist AP | AnnofreeOD · nuScenes val (3-class) | Cyclist AP: 0.0 -> 7.9 | Improved |
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.
- UNION — UNION: Unsupervised 3D Object Detection using Object Appearance-based Pseudo-Classes (NeurIPS 2024) · better on 4 of 4 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
- Zoo3D: Zero-Shot 3D Object Detection at Scene Level (CVPR 2026 · similarity 0.90)
- OpenBox: Annotate Any Bounding Boxes in 3D (NeurIPS 2025 · similarity 0.89)
- ImOV3D: Learning Open Vocabulary Point Clouds 3D Object Detection from Only 2D Images (NeurIPS 2024 · similarity 0.89)
- UNION: Unsupervised 3D Object Detection using Object Appearance-based Pseudo-Classes (NeurIPS 2024 · similarity 0.88)
- Object Detection Benchmarks are Incomplete: The Role of Label Errors and Annotation Uncertainty (NeurIPS 2026 · similarity 0.87)
- WeakPCSOD: Overcoming the Bias of Box Annotations for Weakly Supervised Point Cloud Salient Object Detection (AAAI 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.
