Research map / Point clouds / Segmentation and detection
Sparse2Dense: Learning to Densify 3D Features for 3D Object Detection
Tianyu Wang, Xiaowei Hu, Zhengzhe Liu, Chi-Wing Fu
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 Sparse2Dense to SECOND | SECOND · Waymo Open validation, Vehicle-L2 | mAP: 59.42 -> 63.49 | Improved |
| Add Sparse2Dense to CenterPoint-Pillar | CenterPoint-Pillar · Waymo Open validation, Vehicle-L2 | mAP: 64.12 -> 68.11 | Improved |
| Add Sparse2Dense to CenterPoint | CenterPoint · Waymo Open validation, Vehicle-L2 | mAP: 65.52 -> 68.21 | Improved |
| Add Sparse2Dense to CenterPoint-Voxel | CenterPoint-Voxel · Waymo Domain Adaptation, Vehicle-L2 | mAP: 48.36 -> 51.01 | Improved |
| Add feature and heat-map distillation (ablation) | CenterPoint · Waymo Open validation, Vehicle-L2 | mAP: 63.03 -> 63.84 | Improved |
| Add the S2D module to the distillation variant (ablation) | CenterPoint · Waymo Open validation, Vehicle-L2 | mAP: 63.84 -> 65.75 | Improved |
| Add PCR to the S2D variant (ablation) | CenterPoint · Waymo Open validation, Vehicle-L2 | mAP: 65.75 -> 66.12 | Improved |
| Add PCR to the S2D variant (ablation) | CenterPoint · Waymo Open validation, Pedestrian-L2 | mAP: 67.62 -> 67.47 | Hurt |
| Add PCR to the S2D variant (ablation) | CenterPoint · Waymo Open validation, Cyclist-L2 | mAP: 68.50 -> 68.69 | Improved |
| Add the S2D module to the distillation variant (ablation) | CenterPoint · Waymo Open validation, Pedestrian-L2 | mAP: 67.04 -> 67.62 | Improved |
| Add the S2D module to the distillation variant (ablation) | CenterPoint · Waymo Open validation, Cyclist-L2 | mAP: 67.59 -> 68.50 | Improved |
| Add feature and heat-map distillation (ablation) | CenterPoint · Waymo Open validation, Cyclist-L2 | mAP: 65.03 -> 67.59 | Improved |
Closest papers by meaning
- STD: Sparse-to-Dense 3D Object Detector for Point Cloud (ICCV 2019 · similarity 0.90)
- Point2Seq: Detecting 3D Objects As Sequences (CVPR 2022 · similarity 0.90)
- SparseLIF: High-Performance Sparse LiDAR-Camera Fusion for 3D Object Detection (ECCV 2024 · similarity 0.89)
- Not All Points Are Equal: Learning Highly Efficient Point-Based Detectors for 3D LiDAR Point Clouds (CVPR 2022 · similarity 0.89)
- Voxel or Pillar: Exploring Efficient Point Cloud Representation for 3D Object Detection (AAAI 2024 · similarity 0.89)
- SAFDNet: A Simple and Effective Network for Fully Sparse 3D Object Detection (CVPR 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.
