Research map / Point clouds / Segmentation and detection
PRED: Pre-training via Semantic Rendering on LiDAR Point Clouds
Hao Yang, Haiyang Wang, Di Dai, Liwei Wang
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 PRED pre-training to DSVT-P CenterPoint | CenterPoint detector · nuScenes val | mAP: 61.5 -> 64.2 | Improved |
| add PRED pre-training to Second | Second · nuScenes val | mAP: 52.2 -> 55.2 | Improved |
| add PRED pre-training to DSVT | DSVT · nuScenes test | mAP: 68.4 -> 70.1 | Improved |
| add PRED pre-training to CenterPoint on ONCE | CenterPoint · ONCE val | mAP: 64.28 -> 67.41 | Improved |
| increase pre-training data from small to large (ablation) | PRED CenterPoint · ONCE val | mAP: 65.87 -> 66.72 -> 67.41 | Improved |
| add PRED pre-training to DSVT for BEV map segmentation | DSVT · nuScenes BEV map | mIoU: 51.6 -> 55.0 | Improved |
| swap backbone to 2D Conv (ablation) | PRED · nuScenes val | mAP: 50.0 -> 53.0 | Improved |
| replace direct projection with neural rendering w/ depth (ablation) | PRED · nuScenes val | mAP: 62.0 -> 64.2 | Improved |
| add depth supervision to rendering (ablation) | PRED · nuScenes val | mAP: 63.5 -> 64.2 | Improved |
| add depth filtering to projection (ablation) | PRED projection · nuScenes val | mAP: 62.0 -> 62.8 | Improved |
| add confidence weight and class balance sampling to semantic loss (ablation) | PRED · nuScenes val | mAP: 62.5 -> 64.2 | Improved |
| use lidar view instead of image view for depth supervision (ablation) | PRED · nuScenes val | mAP: 63.6 -> 64.2 | Improved |
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.
- ProposalContrast — ProposalContrast: Unsupervised Pre-training for LiDAR-Based 3D Object Detection (ECCV 2022) · better on 2 of 2 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
- TREND: Unsupervised 3D Representation Learning via Temporal Forecasting for LiDAR Perception (NeurIPS 2025 · similarity 0.89)
- Point Cloud Pre-training with Diffusion Models (CVPR 2024 · similarity 0.89)
- Pre-training LiDAR-based 3D Object Detectors through Colorization (ICLR 2024 · similarity 0.88)
- BEV-MAE: Bird’s Eye View Masked Autoencoders for Point Cloud Pre-training in Autonomous Driving Scenarios (AAAI 2024 · similarity 0.88)
- Visual Point Cloud Forecasting enables Scalable Autonomous Driving (CVPR 2024 · similarity 0.88)
- VLM2Scene: Self-Supervised Image-Text-LiDAR Learning with Foundation Models for Autonomous Driving Scene Understanding (AAAI 2024 · 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.
