Research map / Detection and segmentation / CNN features
Lane Detection Transformer Based on Multi-Frame Horizontal and Vertical Attention and Visual Transformer Module
Han Zhang, Yunchao Gu, Xinliang Wang, Junjun Pan, Minghui Wang
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 multi-frame (5 frames) input (ablation) | LSTR · VIL-100 (ablation) | F1: 0.43% on F1, 0.89% on Accuracy, 3.49% on FP | Improved |
| add multi-frame (5 frames) input (ablation) | Ours · VIL-100 (ablation) | F1: 5.49% on F1, 2.33% improvement on Accuracy, a reduction of 3.43% on FP, and a reduction of 4% on FN | Improved |
| feed 5 identical frames instead of single frame (ablation) | Ours · VIL-100 (ablation) | F1: 79.51 -> 76.85 | Hurt |
| replace Transformer encoder with MHVA module (ablation) | LSTR multi-frame · VIL-100 (ablation) | F1: 70.73 -> 80.54 | Improved |
| replace Transformer decoder with VT modules (ablation) | MHVA model · VIL-100 (ablation) | F1: 80.54 -> 84.59 | Improved |
| use MHVA+VT design instead of LSTR on single frame (ablation) | LSTR · VIL-100 (ablation) | F1: 70.30 -> 79.51 | Improved |
| increase input frames from 3 to 5 (ablation) | Ours · VIL-100 (frame count) | F1: 83.68 -> 84.59 | Improved |
| increase input frames from 5 to 7 (ablation) | Ours · VIL-100 (frame count) | F1: 84.59 -> 84.34 | No clear effect |
| add multi-frame MHVA+VT design over LSTR | LSTR · TuSimple | Accuracy: 96.03 -> 96.17 | No clear effect |
Closest papers by meaning
- A Siamese Transformer with Hierarchical Refinement for Lane Detection (NeurIPS 2024 · similarity 0.89)
- A Hybrid Global-Local Perception Network for Lane Detection (AAAI 2024 · similarity 0.86)
- GSENet:Global Semantic Enhancement Network for Lane Detection (AAAI 2024 · similarity 0.86)
- Sketch and Refine: Towards Fast and Accurate Lane Detection (AAAI 2024 · similarity 0.86)
- DV-3DLane: End-to-end Multi-modal 3D Lane Detection with Dual-view Representation (ICLR 2024 · similarity 0.85)
- GEONet: Global Enhancement and Optimization Network for Lane Detection (AAAI 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.
