Research map / Detection and segmentation / 3D detection and driving
Enhancing Multi-modal Features Using Local Self-Attention for 3D Object Detection
Hao Li, Zehan Zhang, Xian Zhao, Yulong Wang, Yuxi Shen, Shiliang Pu, Hui Mao
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 camera features by naive concatenation fusion to range-only detector (ablation) | EMMF-Det range baseline · Waymo val (20% train) | All AP L1: 72.7 -> 73.3 | Improved |
| add local self-attention feature re-weighting fusion instead of naive concatenation (ablation) | EMMF-Det · Waymo val (20% train) | All AP L1: 73.3 -> 74.3 | Improved |
| add local self-attention feature re-weighting fusion to range-only detector (ablation) | EMMF-Det · Waymo val (20% train) | All AP L2: 64.3 -> 66.4 | Improved |
| increase neighborhood size k from 8 to 16 in fusion candidate sets (ablation) | EMMF-Det · Waymo val (20% train) | All AP L1: 74.3 -> 73.9 | Hurt |
| k=16 still beats naive concatenation (ablation) | EMMF-Det · Waymo val (20% train) | All AP L1: 73.3 -> 73.9 | Improved |
| add multi-modal copy-paste augmentation (ablation) | EMMF-Det · Waymo val (20% train) | All AP L1: 71.3 -> 73.0 | Improved |
| use linear combination of pixels instead of direct pasting in camera copy-paste (ablation) | EMMF-Det · Waymo val (20% train) | All AP L1: 73.0 -> 73.3 | Improved |
| add noise jitter on range images (ablation) | EMMF-Det · Waymo val (20% train) | All AP L1: 73.3 -> 74.3 | Improved |
| add noise jitter on range images (ablation) | EMMF-Det · Waymo val (20% train) | All AP L2: 65.3 -> 66.4 | Improved |
| add IoU prediction module | EMMF-Det · Waymo val (full training) | All AP L1: 76.9 -> 78.8 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- RangeIoUDet — RangeIoUDet: Range Image Based Real-Time 3D Object Detector Optimized by Intersection Over Union (CVPR 2021) · better on 2 of 2 shared results
- RangeDet — RangeDet: In Defense of Range View for LiDAR-Based 3D Object Detection (ICCV 2021) · better on 1 of 1 shared result
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
- CCF: Complementary Collaborative Fusion for Domain Generalized Multi-Modal 3D Object Detection (CVPR 2026 · similarity 0.88)
- 3D-CVF: Generating Joint Camera and LiDAR Features Using Cross-View Spatial Feature Fusion for 3D Object Detection (ECCV 2020 · similarity 0.88)
- MGAF: LiDAR-Camera 3D Object Detection With Multiple Guidance and Adaptive Fusion (TPAMI 2025 · similarity 0.88)
- EPNet++: Cascade Bi-Directional Fusion for Multi-Modal 3D Object Detection (TPAMI 2022 · similarity 0.88)
- PromptDet: A Lightweight 3D Object Detection Framework with LiDAR Prompts (AAAI 2025 · similarity 0.86)
- SimDistill: Simulated Multi-Modal Distillation for BEV 3D Object Detection (AAAI 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.
