Research map / Detection and segmentation / 3D detection and driving
Temporal Consistent 3D LiDAR Representation Learning for Semantic Perception in Autonomous Driving
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 |
|---|---|---|---|
| Pre-train with temporal object association | MinkUNet for semantic segmentation · SemanticKITTI val, 10% labels | mIoU: 58.68 -> 60.34 | Helped |
| Pre-train with temporal object association | MinkUNet for semantic segmentation · SemanticKITTI val, 100% labels | mIoU: 61.45 -> 61.47 | No clear effect |
| Pre-train with temporal object association | MinkUNet for semantic segmentation · nuScenes val mini | mIoU: 38.39 -> 39.36 | Helped |
| Pre-train with temporal object association | MinkUNet for semantic segmentation · SemanticKITTI linear evaluation | mIoU: 27.42 -> 35.90 | Helped |
| Pre-train with temporal object association | MinkUNet for panoptic segmentation · SemanticKITTI val, 10% labels | PQ: 47.57 -> 51.27 | Helped |
| Pre-train with temporal object association | MinkUNet for panoptic segmentation · SemanticKITTI val, 100% labels | PQ: 56.73 -> 56.57 | Hurt |
| Pre-train with temporal object association | MinkUNet for panoptic segmentation · nuScenes val mini | PQ: 28.84 -> 32.22 | Helped |
| Pre-train with temporal object association | MinkUNet for KITTI object detection · KITTI car moderate | AP50: 82.13 -> 82.15 | No clear effect |
| Pre-train with temporal object association | MinkUNet for KITTI object detection · KITTI pedestrian hard | AP50: 57.71 -> 58.74 | Helped |
| Pre-train with temporal object association | MinkUNet for KITTI object detection · KITTI cyclist moderate | AP50: 75.20 -> 75.44 | Helped |
| Use 10% of labeled scans with TARL pre-training instead of training from scratch with all labels | MinkUNet semantic segmentation · SemanticKITTI | mIoU: 59.03 -> 60.34 | Helped |
| Use temporal views of the same object as positive pairs | TARL pre-training · downstream tasks | mIoU | Helped |
Closest papers by meaning
- VLM2Scene: Self-Supervised Image-Text-LiDAR Learning with Foundation Models for Autonomous Driving Scene Understanding (AAAI 2024 · similarity 0.91)
- HiLoTs: High-Low Temporal Sensitive Representation Learning for Semi-Supervised LiDAR Segmentation in Autonomous Driving (CVPR 2025 · similarity 0.91)
- TREND: Unsupervised 3D Representation Learning via Temporal Forecasting for LiDAR Perception (NeurIPS 2025 · similarity 0.90)
- Multi-Space Alignments Towards Universal LiDAR Segmentation (CVPR 2024 · similarity 0.88)
- Self-Supervised Class-Agnostic Motion Prediction with Spatial and Temporal Consistency Regularizations (CVPR 2024 · similarity 0.88)
- Exploring Scene Affinity for Semi-Supervised LiDAR Semantic Segmentation (CVPR 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.