Research map / Point clouds / Segmentation and detection
3D Object Detection Using Scale Invariant and Feature Reweighting Networks
Read the paperCheck 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 |
|---|---|---|---|
| Replace the three SA modules with SENet (ablation) | F-PointNet (V2) · KITTI validation | 3D mAP (%): 65.58 -> 65.73 | Helped |
| Add PointSIFT (ablation) | V2-SENet · KITTI validation | 3D mAP (%): 65.73 -> 66.18 | Helped |
| Add image features and RGB information (ablation) | V2-SENet-PointSIFT · KITTI validation | 3D mAP (%): 66.18 -> 66.64 | Helped |
| Fine-tune using the redefined angle loss (ablation) | V2-SENet-PointSIFT-rgb-image · KITTI validation | 3D mAP (%): 66.64 -> 66.99 | Helped |
| Use the proposed angle loss instead of the original loss (ablation) | SIFRNet · KITTI validation | 3D mAP (%): 65.37 -> 65.61; 65.82 -> 66.25; 66.54 -> 66.91; 66.68 -> 66.99 | Helped |
| Use SIFRNet instead of F-PointNet (V1) | 3D detection on SUN-RGBD · SUN-RGBD test, 32 input points | 3D mAP: 43.9 -> 51.1 | Helped |
| Use SIFRNet instead of F-PointNet (V1) | 3D detection on SUN-RGBD · SUN-RGBD test, 2048 input points | 3D mAP: 56.1 -> 58.4 | Helped |
| Improve 2D detector performance while holding the comparison setup | SIFRNet · SUN-RGBD test | 3D mAP: 46.1 -> 48.1; 49.6 -> 51.6; 52.2 -> 54.3; 55.5 -> 57.6; 56.1 -> 58.4; 84.1 -> 86.7 | Helped |
Closest papers by meaning
- Frustum PointNets for 3D Object Detection From RGB-D Data (CVPR 2018 · similarity 0.90)
- TANet: Robust 3D Object Detection from Point Clouds with Triple Attention (AAAI 2020 · similarity 0.88)
- Uni3DETR: Unified 3D Detection Transformer (NeurIPS 2023 · similarity 0.88)
- Structure Aware Single-Stage 3D Object Detection From Point Cloud (CVPR 2020 · similarity 0.88)
- Density-Based Clustering for 3D Object Detection in Point Clouds (CVPR 2020 · similarity 0.88)
- D3Feat: Joint Learning of Dense Detection and Description of 3D Local Features (CVPR 2020 · 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.