Research map / Detection and segmentation / 3D detection and driving
3D-MAN: 3D Multi-Frame Attention Network for Object Detection
Check 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 |
|---|---|---|---|
| Use 16-frame multi-view inputs rather than single-frame input | 3D-MAN · Waymo validation, Vehicle LEVEL_1 | 3D AP: 69.03 -> 74.53 | Helped |
| Use Hungarian matching instead of mask assignment (ablation) | 3D-MAN FSD · Waymo validation, Vehicle LEVEL_1, mini-validation | AP: 44.5 -> 64.8 | Helped |
| Use Hungarian matching instead of centeredness assignment (ablation) | 3D-MAN FSD · Waymo validation, Vehicle LEVEL_1, mini-validation | AP: 63.7 -> 64.8 | Helped |
| Fuse multi-frame information using point concatenation (ablation) | 3D-MAN · Waymo validation, Vehicle LEVEL_1, mini-validation | AP: 68.2 -> 70.5 | Helped |
| Replace the single-frame baseline with self-attention over past proposals (ablation) | 3D-MAN · Waymo validation, Vehicle LEVEL_1, mini-validation | AP: 68.2 -> 70.1 | Helped |
| Use MVAA instead of point concatenation (ablation) | 3D-MAN · Waymo validation, Vehicle LEVEL_1, mini-validation | AP: 70.5 -> 72.5 | Helped |
| Use MVAA instead of relation self-attention (ablation) | 3D-MAN · Waymo validation, Vehicle LEVEL_1, mini-validation | AP: 70.1 -> 72.5 | Helped |
| Add cross-view loss rather than using no auxiliary alignment loss (ablation) | 3D-MAN MVAA · Waymo validation, Vehicle LEVEL_1, mini-validation | AP: 70.7 -> 72.5 | Helped |
| Use cross-view loss instead of correspondence loss (ablation) | 3D-MAN MVAA · Waymo validation, Vehicle LEVEL_1, mini-validation | AP: 71.4 -> 72.5 | Helped |
| Increase the number of input frames from 1 to 16 (ablation) | 3D-MAN · Waymo validation, Vehicle LEVEL_1, mini-validation | AP: 68.2 -> 74.3 | Helped |
| Increase input frames from 4 to 16 (ablation) | 3D-MAN · Waymo validation, Vehicle velocity breakdown | Vehicle AP: 69.5 -> 73.2; 68.6 -> 70.4; 67.1 -> 68.9; 78.3 -> 79.2 | Helped |
Closest papers by meaning
- Time3D: End-to-End Joint Monocular 3D Object Detection and Tracking for Autonomous Driving (CVPR 2022 · similarity 0.88)
- Omni3D: A Large Benchmark and Model for 3D Object Detection in the Wild (CVPR 2023 · similarity 0.88)
- Towards Generalizable Multi-Camera 3D Object Detection via Perspective Rendering (AAAI 2025 · similarity 0.86)
- ImOV3D: Learning Open Vocabulary Point Clouds 3D Object Detection from Only 2D Images (NeurIPS 2024 · similarity 0.86)
- UNION: Unsupervised 3D Object Detection using Object Appearance-based Pseudo-Classes (NeurIPS 2024 · similarity 0.86)
- ImageNet3D: Towards General-Purpose Object-Level 3D Understanding (NeurIPS 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.