Research map / Point clouds / Segmentation and detection
EPNet++: Cascade Bi-Directional Fusion for Multi-Modal 3D Object Detection
Zhe Liu, Tengteng Huang, Bingling Li, Xiwu Chen, Xi Wang, Xiang Bai
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 CB-Fusion only (no MC loss) (ablation) | EPNet LiDAR-only · KITTI val Cars Moderate (16-beam) | AP: 63.30 -> 68.83 | Improved |
| add LI-Fusion only (ablation) | EPNet LiDAR-only · KITTI val Cars Moderate (16-beam) | AP: 63.30 -> 67.72 | Improved |
| add MC loss only (ablation) | EPNet LiDAR-only · KITTI val Cars Moderate (16-beam) | AP: 63.30 -> 63.69 | Improved |
| add MC loss on top of CB-Fusion (ablation) | EPNet++ · KITTI val Cars Moderate (16-beam) | AP: 68.83 -> 69.12 | Improved |
| add attention to CB-Fusion on 16-beam Pedestrians (ablation) | EPNet++ · KITTI val 16-beam Pedestrians | AP: 3.45% | Improved |
| add attention to LI-Fusion on 16-beam Pedestrians (ablation) | EPNet · KITTI val 16-beam Pedestrians | AP: 0.90% | Improved |
| use cascade CB-Fusion instead of parallel bi-directional fusion (ablation) | EPNet++ · KITTI val (64-beam) | AP: 73.25 -> 74.04 | Improved |
| use CB-Fusion instead of reversed order (ablation) | EPNet++ · KITTI val (64-beam) | AP: 73.39 -> 74.04 | Improved |
| add CB-Fusion to LiDAR-only EPNet vs PointPainting fusion (ablation) | EPNet baseline · KITTI val (64-beam) | AP: 72.14 -> 73.11 | Improved |
| use transformer-based CB-Fusion instead of attention-gated CB-Fusion (ablation) | EPNet++ · KITTI val (64-beam) | AP: 73.11 -> 72.50 | Hurt |
| add LI-Fusion without I2P attention gate (ablation) | LiDAR-only baseline · KITTI val (noisy-data analysis) | AP: 67.89 -> 67.77 | Hurt |
| add I2P attention gate to LI-Fusion (ablation) | LI-Fusion · KITTI val (noisy-data analysis) | AP: 67.77 -> 69.63 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 5 of 5 on most shared results.
- EPNet — EPNet: Enhancing Point Features with Image Semantics for 3D Object Detection (ECCV 2020) · better on 5 of 5 shared results
- ImVoteNet* — ImVoteNet: Boosting 3D Object Detection in Point Clouds With Image Votes (CVPR 2020) · better on 1 of 1 shared result
- PV-RCNN — PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection (CVPR 2020) · better on 1 of 1 shared result
- TANet — TANet: Robust 3D Object Detection from Point Clouds with Triple Attention (AAAI 2020) · better on 1 of 1 shared result
- PointRCNN — PointRCNN: 3D Object Proposal Generation and Detection From Point Cloud (CVPR 2019) · 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
- EPNet: Enhancing Point Features with Image Semantics for 3D Object Detection (ECCV 2020 · similarity 0.92)
- LoGoNet: Towards Accurate 3D Object Detection With Local-to-Global Cross-Modal Fusion (CVPR 2023 · similarity 0.91)
- CCF: Complementary Collaborative Fusion for Domain Generalized Multi-Modal 3D Object Detection (CVPR 2026 · similarity 0.91)
- ViKIENet: Towards Efficient 3D Object Detection with Virtual Key Instance Enhanced Network (CVPR 2025 · similarity 0.89)
- SimDistill: Simulated Multi-Modal Distillation for BEV 3D Object Detection (AAAI 2024 · similarity 0.88)
- Voxel or Pillar: Exploring Efficient Point Cloud Representation for 3D Object Detection (AAAI 2024 · similarity 0.88)
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.
