Research map / Detection and segmentation / CNN features
Spatial as Deep: Spatial CNN for Traffic Scene Understanding
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 |
|---|---|---|---|
| add SCNN (slice-by-slice message passing) to baseline | LargeFOV · CULane | F1 (IoU 0.5): 63.2 -> 71.6 | Helped |
| add an extra conv layer (parameter-matched control) (ablation) | LargeFOV · CULane | F1 (IoU 0.5): 63.2 -> 64.0 | No clear effect |
| add more propagation directions (D -> DURL) (ablation) | SCNN · CULane | F1 (IoU 0.5): 68.6 -> 70.4 | Helped |
| increase kernel width from 1 to 9 (ablation) | SCNN · CULane | F1 (IoU 0.5): 66.3 -> 71.6 | Helped |
| apply SCNN at top hidden layer instead of output (ablation) | SCNN · CULane | F1 (IoU 0.5): 68.8 -> 71.6 | Helped |
| use sequential instead of parallel message passing (ablation) | SCNN · CULane | F1 (IoU 0.5): 65.2 -> 71.6 | Helped |
| add DenseCRF post-processing | LargeFOV · CULane | F1 (IoU 0.5): 63.2 -> 61.0 | Hurt |
| add ReNet LSTM layers | LargeFOV · CULane | F1 (IoU 0.5): 63.2 -> 62.9 | No clear effect |
| add SCNN | ResNet-101 · Cityscapes val | mIoU: 75.6 -> 76.4 | Helped |
| replace dense CRF with SCNN | message passing · CPU runtime | runtime (ms): 737 -> 176 | Helped |
Closest papers by meaning
- Sketch and Refine: Towards Fast and Accurate Lane Detection (AAAI 2024 · similarity 0.88)
- Laneformer: Object-Aware Row-Column Transformers for Lane Detection (AAAI 2022 · similarity 0.88)
- GEONet: Global Enhancement and Optimization Network for Lane Detection (AAAI 2025 · similarity 0.87)
- Spatio-Temporal Graph Structure Learning for Traffic Forecasting (AAAI 2020 · similarity 0.87)
- GSENet:Global Semantic Enhancement Network for Lane Detection (AAAI 2024 · similarity 0.87)
- LaneSegNet: Map Learning with Lane Segment Perception for Autonomous Driving (ICLR 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.