Research map / Detection and segmentation / 3D detection and driving
HG-Lane: High-Fidelity Generation of Lane Scenes under Adverse Weather and Lighting Conditions without Re-annotation
Daichao Zhao, Qiupu Chen, Feng He, Xin Ning, Qiankun Li
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 HG-Lane generated adverse-weather data (same total size) | CLRNet ResNet-18 · HG-Lane benchmark (CLRNet ResNet-18) | F1@50: 68.89 -> 88.49 | Improved |
| add HG-Lane generated data | CLRNet ResNet-18 · HG-Lane benchmark (CLRNet ResNet-18) | mF1: 45.42 -> 63.03 | Improved |
| add HG-Lane generated data | CLRNet ResNet-18 · HG-Lane benchmark (CLRNet ResNet-18) | F1@50 snow: 51.79 -> 84.88 | Improved |
| scale normal-only training set from 4k to 4k*6 | CLRNet ResNet-18 · HG-Lane benchmark (CLRNet ResNet-18) | F1@50: 68.74 -> 68.89 | No clear effect |
| add lane-label mask to the Canny control map (ablation) | HG-Lane · HG-Lane ablation (CLRNet ResNet-18) | F1@50: 86.81 -> 88.49 | Improved |
| apply Canny before iP2P instead of the reverse order (ablation) | HG-Lane · HG-Lane ablation (CLRNet ResNet-18) | F1@50: 62.56 -> 88.49 | Improved |
| add Canny stage to iP2P-only generation (ablation) | HG-Lane · HG-Lane ablation (CLRNet ResNet-18) | F1@50: 59.75 -> 88.49 | Improved |
| add iP2P stage to Canny-only generation (ablation) | HG-Lane · HG-Lane ablation (CLRNet ResNet-18) | F1@50: 64.01 -> 88.49 | Improved |
| use HG-Lane instead of CycleGAN for generation | CLRNet ResNet-18 · Generation quality comparison (CLRNet ResNet-18 F1@50 and FID) | FID: 99.8 -> 80.1 | Improved |
| use HG-Lane instead of Stable Diffusion for generation | CLRNet ResNet-18 · Generation quality comparison (CLRNet ResNet-18 F1@50 and FID) | F1@50: 69.02 -> 88.49 | Improved |
| train on HG-Lane data instead of CULane | CLRNet ResNet-18 · Real-world test (CLRNet ResNet-18) | F1@50: 53.03 -> 68.54 | Improved |
| train on HG-Lane data instead of CULane | CLRNet ResNet-18 · Real-world night | F1@50: 22.47 -> 60.02 | Improved |
Closest papers by meaning
- FastDraw: Addressing the Long Tail of Lane Detection by Adapting a Sequential Prediction Network (CVPR 2019 · similarity 0.87)
- A Hybrid Global-Local Perception Network for Lane Detection (AAAI 2024 · similarity 0.86)
- GEONet: Global Enhancement and Optimization Network for Lane Detection (AAAI 2025 · similarity 0.86)
- GSENet:Global Semantic Enhancement Network for Lane Detection (AAAI 2024 · similarity 0.86)
- A Siamese Transformer with Hierarchical Refinement for Lane Detection (NeurIPS 2024 · similarity 0.85)
- Rethinking Lanes and Points in Complex Scenarios for Monocular 3D Lane Detection (CVPR 2025 · 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.
