Research map / Detection and segmentation / 3D detection and driving
LinK: Linear Kernel for LiDAR-Based 3D Perception
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 CenterPoint's backbone with a LinK-based backbone | CenterPoint · nuScenes test (LiDAR) | NDS: 65.5 -> 71.0 | Helped |
| Replace CenterPoint's backbone with a LinK-based backbone | CenterPoint · nuScenes val | mAP: 59.0 -> 63.6 | Helped |
| Replace Mink's encoder with a LinK-based encoder | Mink · SemanticKITTI test | mIoU: 68.0 -> 70.7 | Helped |
| Add LinK to CenterPoint | CenterPoint · Waymo Detection val | mAPH: 63.4 -> 64.6 | Helped |
| Add LinK to CenterPoint | CenterPoint · KITTI Detection val | mAP (Moderate): 61.7 -> 63.8 | Helped |
| Add LinK to Mink | Mink · nuScenes Segmentation val | mIoU: 74.8 -> 76.4 | Helped |
| Use LinK instead of Conv 7×7×7 in the large-kernel branch, retaining the bypass and ResBlock (ablation) | LinK backbone · SemanticKITTI val | mIoU: 66.8 -> 67.5 | Helped |
| Add the 3×3×3 bypass branch (ablation) | LinK backbone · SemanticKITTI val | mIoU: 65.6 -> 67.5 | Helped |
| Add the ResBlock branch (ablation) | LinK backbone · SemanticKITTI val | mIoU: 65.7 -> 67.5 | Helped |
| Add learnable frequency (ablation) | LinK, r×s=2×3 · SemanticKITTI val | mIoU: 67.3 -> 67.5 | Helped |
| Add group sharing (ablation) | LinK, r×s=2×3 · SemanticKITTI val | mIoU: 67.3 -> 67.1 | Hurt |
| Add group sharing (ablation) | LinK, r×s=3×5 · SemanticKITTI val | mIoU: 66.2 -> 67.5 | Helped |
Closest papers by meaning
- LSK3DNet: Towards Effective and Efficient 3D Perception with Large Sparse Kernels (CVPR 2024 · similarity 0.92)
- Fully Sparse 3D Object Detection (NeurIPS 2022 · similarity 0.90)
- LidarMultiNet: Towards a Unified Multi-Task Network for LiDAR Perception (AAAI 2023 · similarity 0.90)
- LargeKernel3D: Scaling Up Kernels in 3D Sparse CNNs (CVPR 2023 · similarity 0.89)
- Point Density-Aware Voxels for LiDAR 3D Object Detection (CVPR 2022 · similarity 0.89)
- TREND: Unsupervised 3D Representation Learning via Temporal Forecasting for LiDAR Perception (NeurIPS 2025 · 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.