Research map / 3D vision / 3D generation
LetsMap: Unsupervised Representation Learning for Label-Efficient Semantic BEV Mapping
Nikhil Gosala, Kürsat Petek, B Ravi Kiran, Senthil Yogamani, Paulo L. J. Drews-Jr, Wolfram Burgard, Abhinav Valada
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 unsupervised pretraining before BEV finetuning (ablation) | LetsMap · KITTI-360 (1% BEV labels) | mIoU: 26.47 -> 28.96 | Improved |
| add unsupervised pretraining with 100% BEV labels (ablation) | LetsMap · KITTI-360 (100% BEV labels) | mIoU: 32.68 -> 32.88 | No clear effect |
| add geometric pathway (implicit field) to pretraining (ablation) | LetsMap · KITTI-360 (1% BEV labels) | mIoU: 26.47 -> 27.75 | Improved |
| add temporal masked autoencoder semantic pathway (ablation) | LetsMap with geometric pathway · KITTI-360 (1% BEV labels) | mIoU: 27.75 -> 28.96 | Improved |
| increase masking ratio to 75% (versus 0%, 25%, 50%) (ablation) | T-MAE pretraining · KITTI-360 (1% BEV labels) | mIoU: 27.75 -> 28.96 | Improved |
| increase masking ratio to 90% (ablation) | T-MAE pretraining · KITTI-360 (1% BEV labels) | mIoU: 28.96 -> 27.31 | Hurt |
| use LetsMap instead of fully supervised PoBEV at 1% BEV labels | BEV mapping · KITTI-360 (1% BEV labels) | mIoU: 19.09 -> 28.96 | Improved |
| use unsupervised LetsMap on nuScenes versus fully supervised PoBEV at 100% labels | BEV mapping · nuScenes | mIoU: 30.90 -> 27.66 | Hurt |
| LetsMap vs SkyEye with FV labels on dynamic classes | BEV mapping · KITTI-360 car and truck | IoU: Car 39.77 (PoBEV) vs 30.94 | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- SkyEye — SkyEye: Self-Supervised Bird's-Eye-View Semantic Mapping Using Monocular Frontal View Images (CVPR 2023) · better on 1 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
- Improving Bird's Eye View Semantic Segmentation by Task Decomposition (CVPR 2024 · similarity 0.91)
- Map It Anywhere: Empowering BEV Map Prediction using Large-scale Public Datasets (NeurIPS 2024 · similarity 0.89)
- VQ-Map: Bird's-Eye-View Map Layout Estimation in Tokenized Discrete Space via Vector Quantization (NeurIPS 2024 · similarity 0.88)
- A Black-Box Evaluation Framework for Semantic Robustness in Bird’s Eye View Detection (AAAI 2025 · similarity 0.87)
- BEV-TSR: Text-Scene Retrieval in BEV Space for Autonomous Driving (AAAI 2025 · similarity 0.87)
- M-BEV: Masked BEV Perception for Robust Autonomous Driving (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.
