Research map / Detection and segmentation / 3D detection and driving
Monocular BEV Perception of Road Scenes via Front-to-Top View Projection
Wenxi Liu, Qi Li, Weixiang Yang, Jiaxin Cai, Yuanlong Yu, Yuexin Ma, Shengfeng He, Jia Pan
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 MLP to baseline encoder-decoder (ablation) | baseline · KITTI 3D Object | mIoU (%): 22.31 -> 27.42 | Improved |
| add cross-view correlation (ablation) | FTVP network · KITTI 3D Object | mIoU (%): 27.42 -> 35.03 | Improved |
| add cycle structure and cycle loss (ablation) | FTVP network · KITTI 3D Object | mAP (%): 46.33 -> 47.29 | Improved |
| add feature selection (ablation) | FTVP network · KITTI 3D Object | mAP (%): 47.29 -> 54.53 | Improved |
| add multi-scale FTVPs (ablation) | FTVP network · KITTI 3D Object | mIoU (%): 39.97 -> 37.83; mAP 54.53 -> 55.02 | Mixed |
| add deep supervision to multi-scale model (ablation) | FTVP network · KITTI 3D Object | mAP (%): 55.02 -> 59.05 | Improved |
| add deep supervision to single-FTVP model (ablation) | single FTVP · KITTI 3D Object | mIoU (%): 39.48 vs 39.97 | No clear effect |
| use X as key and X'' as value instead of X' for both (ablation) | CVT · KITTI 3D Object | mAP (%): 53.20 -> 59.05 | Improved |
| use 3 FTVPs instead of naive convolution (ablation) | Ours · KITTI 3D Object | mAP (%): 58.40 -> 59.05 | Improved |
| use 3 FTVPs instead of 4 (ablation) | Ours · KITTI 3D Object | mAP (%): 57.70 -> 59.05 | Improved |
| use multi-scale FTVP model instead of CVT preliminary version | Ours · KITTI Odometry road layout | mAP (%): 86.69 -> 87.82 | Improved |
Closest papers by meaning
- Projecting Your View Attentively: Monocular Road Scene Layout Estimation via Cross-View Transformation (CVPR 2021 · similarity 0.95)
- LiDAR2Map: In Defense of LiDAR-Based Semantic Map Construction Using Online Camera Distillation (CVPR 2023 · similarity 0.87)
- SafeMap: Robust HD Map Construction from Incomplete Observations (ICML 2025 · similarity 0.87)
- Map It Anywhere: Empowering BEV Map Prediction using Large-scale Public Datasets (NeurIPS 2024 · similarity 0.86)
- Effective Multi-sensor Conditioning for Street-view Novel-view Synthesis (NeurIPS 2026 · similarity 0.86)
- Improving Bird's Eye View Semantic Segmentation by Task Decomposition (CVPR 2024 · similarity 0.85)
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.
