Research map / Detection and segmentation / 3D detection and driving
To the Point: Efficient 3D Object Detection in the Range Image With Graph Convolution Kernels
Check 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 |
|---|---|---|---|
| add camera features at the input to the first extractor layer (ablation) | PPC + EdgeConv · Waymo Open Dataset validation, pedestrians | 3D APH_L2: 59.6 -> 61.5 | Helped |
| replace Conv2D with EdgeConv kernels throughout (ablation) | PPC · Waymo Open Dataset validation, pedestrians | 3D APH_L2: 47.0 -> 59.6 | Helped |
| replace Conv2D with PointNet kernels throughout (ablation) | PPC · Waymo Open Dataset validation, pedestrians | 3D APH_L2: 47.0 -> 57.9 | Helped |
| replace Conv2D with RQ-Conv2D kernels throughout (ablation) | PPC · Waymo Open Dataset validation, pedestrians | 3D APH_L2: 47.0 -> 54.3 | Helped |
| use self-attention kernels instead of Conv2D (ablation) | PPC · Waymo Open Dataset validation, pedestrians | 3D APH_L2: 47.0 -> 43.7 | Hurt |
| replace Conv2D with EdgeConv kernels throughout (ablation) | PPC · Waymo Open Dataset validation, vehicles | 3D APH_L2: 52.2 -> 56.7 | Helped |
| replace Conv2D with PointNet kernels throughout (ablation) | PPC · Waymo Open Dataset validation, vehicles | 3D APH_L2: 52.2 -> 56.2 | Helped |
| replace Conv2D with RQ-Conv2D kernels throughout (ablation) | PPC · Waymo Open Dataset validation, vehicles | 3D APH_L2: 52.2 -> 49.2 | Hurt |
| replace Conv2D with EdgeConv in FE1 (ablation) | PPC backbone · Waymo Open Dataset validation, pedestrians | 3D APH_L2: 47.0 -> 56.8 | Helped |
| replace Conv2D with EdgeConv in FE2 (ablation) | PPC backbone · Waymo Open Dataset validation, pedestrians | 3D APH_L2: 47.0 -> 54.9 | Helped |
| replace Conv2D with EdgeConv in FE3 (ablation) | PPC backbone · Waymo Open Dataset validation, pedestrians | 3D APH_L2: 47.0 -> 51.7 | Helped |
| replace Conv2D with EdgeConv in FA (ablation) | PPC backbone · Waymo Open Dataset validation, pedestrians | 3D APH_L2: 47.0 -> 53.2 | Helped |
Closest papers by meaning
- LargeKernel3D: Scaling Up Kernels in 3D Sparse CNNs (CVPR 2023 · similarity 0.89)
- PerspectiveNet: 3D Object Detection from a Single RGB Image via Perspective Points (NeurIPS 2019 · similarity 0.89)
- 3D-MAN: 3D Multi-Frame Attention Network for Object Detection (CVPR 2021 · similarity 0.88)
- HVPR: Hybrid Voxel-Point Representation for Single-Stage 3D Object Detection (CVPR 2021 · similarity 0.87)
- Convolution in the Cloud: Learning Deformable Kernels in 3D Graph Convolution Networks for Point Cloud Analysis (CVPR 2020 · similarity 0.87)
- Towards Generalizable Multi-Camera 3D Object Detection via Perspective Rendering (AAAI 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.