Research map / Point clouds / Segmentation and detection
Improving 3D Object Detection With Channel-Wise Transformer
Hualian Sheng, Sijia Cai, Yuan Liu, Bing Deng, Jianqiang Huang, Xian-Sheng Hua, Min-Jian Zhao
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 self-attention encoding (ablation) | CT3D · KITTI val car 3D (40 recall) | AP moderate: 79.20 -> 85.10 | Improved |
| add keypoints subtraction in proposal-to-point embedding (ablation) | CT3D · KITTI val car 3D (40 recall) | AP moderate: 83.41 -> 85.10 | Improved |
| replace standard decoding with extended channel-wise re-weighting (ablation) | CT3D · KITTI val car 3D (40 recall) | AP moderate: 85.10 -> 85.82 | Improved |
| replace channel-wise re-weighting with extended channel-wise re-weighting (ablation) | CT3D · KITTI val car 3D (40 recall) | AP moderate: 85.34 -> 85.82 | Improved |
| replace standard decoding with plain channel-wise re-weighting (ablation) | CT3D · KITTI val car 3D (40 recall) | AP moderate: 85.10 -> 85.34 | Improved |
| add CT3D two-stage refinement on SECOND RPN (ablation) | SECOND · KITTI val car 3D (40 recall) | AP moderate: 80.35 -> 85.82 | Improved |
| add CT3D two-stage refinement on PointPillar RPN (ablation) | PointPillar · KITTI val car 3D (40 recall) | AP moderate: 79.26 -> 84.08 | Improved |
| use CT3D instead of Voxel-RCNN as refinement (val, 11 recall) | CT3D · KITTI val car 3D (11 recall) | AP moderate: 84.52 -> 86.06 | Improved |
| use CT3D instead of Voxel-RCNN (Waymo LEVEL_1 3D) | CT3D · Waymo val vehicle 3D LEVEL_1 | AP moderate: 75.59 -> 76.30 | Improved |
| use CT3D instead of Voxel-RCNN (Waymo LEVEL_2 3D) | CT3D · Waymo val vehicle 3D LEVEL_2 | AP moderate: 66.59 -> 69.04 | Improved |
| use CT3D instead of Voxel-RCNN on KITTI test easy level | CT3D · KITTI test car 3D (40 recall) | AP easy: 90.90 -> 87.83 | Hurt |
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.
- PV-RCNN — PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection (CVPR 2020) · 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
- CT3D++: Improving 3D Object Detection with Keypoint-Induced Channel-wise Transformer (IJCV 2025 · similarity 0.96)
- An End-to-End Transformer Model for 3D Object Detection (ICCV 2021 · similarity 0.90)
- 3D Object Detection With Pointformer (CVPR 2021 · similarity 0.89)
- 3DPPE: 3D Point Positional Encoding for Transformer-based Multi-Camera 3D Object Detection (ICCV 2023 · similarity 0.89)
- ImOV3D: Learning Open Vocabulary Point Clouds 3D Object Detection from Only 2D Images (NeurIPS 2024 · similarity 0.87)
- State Space Model Meets Transformer: A New Paradigm for 3D Object Detection (ICLR 2025 · 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.
