Research map / Detection and segmentation / CNN features
Recognizing Vector Graphics without Rasterization
XINYANG JIANG, Lu Liu, Caihua Shan, Yifei Shen, Xuanyi Dong, Dongsheng 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 position-wise edges (ablation) | YOLaT · SESYD Floorplans test | mAP: 87.17 -> 90.59 | Improved |
| add stroke-wise edges (ablation) | YOLaT · SESYD Floorplans test | mAP: 86.00 -> 90.59 | Improved |
| add edge attributes (ablation) | YOLaT · SESYD Floorplans test | mAP: 86.25 -> 90.59 | Improved |
| include each node's own representation in the stroke-wise update (ablation) | YOLaT stroke-wise stream · SESYD Floorplans test | mAP: 88.40 -> 90.59 | Improved |
| include neighbor-minus-node feature differences in the stroke-wise update (ablation) | YOLaT stroke-wise stream · SESYD Floorplans test | mAP: 87.83 -> 90.59 | Improved |
| aggregate position-wise features only at the last GNN layer rather than at every layer (ablation) | YOLaT position-wise stream · SESYD Floorplans test | mAP: 87.90 -> 90.59 | Improved |
| add neighbor-minus-node feature differences to the position-wise stream (ablation) | YOLaT position-wise stream · SESYD Floorplans test | mAP: 90.59 -> 90.39 | No clear effect |
| use the proposed aggregation function instead of GCN aggregation (ablation) | YOLaT GNN · SESYD Floorplans test | mAP: 83.32 -> 90.59 | Improved |
| use the proposed aggregation function instead of GAT aggregation (ablation) | YOLaT GNN · SESYD Floorplans test | mAP: 83.92 -> 90.59 | Improved |
| use the proposed aggregation function instead of GraphSage aggregation (ablation) | YOLaT GNN · SESYD Floorplans test | mAP: 85.26 -> 90.59 | Improved |
Closest papers by meaning
- Optimize & Reduce: A Top-Down Approach for Image Vectorization (AAAI 2024 · similarity 0.83)
- VFIG: Vectorizing Complex Figures in SVG with Vision-Language Models (NeurIPS 2026 · similarity 0.83)
- StrokeNUWA—Tokenizing Strokes for Vector Graphic Synthesis (ICML 2024 · similarity 0.82)
- VectorArk: Learning Practical Image Vectorization with Rounded Polygon Representation (CVPR 2026 · similarity 0.81)
- Chat2SVG: Vector Graphics Generation with Large Language Models and Image Diffusion Models (CVPR 2025 · similarity 0.81)
- Less is More: Efficient Image Vectorization with Adaptive Parameterization (CVPR 2025 · similarity 0.81)
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.
