Research map / Image generation / Image restoration
Deep Vectorization of Technical Drawings
Vage Egiazarian, Oleg Voynov, Alexey Artemov, Denis Volkhonskiy, Aleksandr Safin, Maria Taktasheva, Denis Zorin, Evgeny Burnaev
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 refinement optimization after the primitive extraction network (ablation) | Ours (NN) · ABC | IoU (%): 65 -> 91 | Improved |
| add refinement optimization after the primitive extraction network (ablation) | Ours (NN) · ABC | Hausdorff distance dH (px): 52 -> 19 | Improved |
| add refinement optimization after the primitive extraction network (ablation) | Ours (NN) · ABC | Mean minimal deviation dM (px): 1.4 -> 0.3 | Improved |
| add refinement optimization after the primitive extraction network (ablation) | Ours (NN) · ABC | Number of primitives: 309 -> 240 | Improved |
| add postprocessing merging of primitives (ablation) | Ours (NN + Refinement) · ABC | Number of primitives: 240 -> 97 | Improved |
| add postprocessing merging of primitives (ablation) | Ours (NN + Refinement) · ABC | IoU (%): 91 -> 77 | Hurt |
| add postprocessing merging of primitives (ablation) | Ours (NN + Refinement) · ABC | Mean minimal deviation dM (px): 0.3 -> 0.6 | Hurt |
| use own preprocessing network instead of MS | preprocessing · DLD preprocessing | IoU (%): 49 -> 92 | Improved |
| use own preprocessing network instead of MS | preprocessing · DLD preprocessing | PSNR: 15.7 -> 25.5 | Improved |
| use full system instead of PVF | vectorization on ABC · ABC | IoU (%): 89 -> 77 | Hurt |
| use full system instead of PVF (primitive count) | vectorization on PFP · PFP | Number of primitives #P: 38k -> 1331 | Improved |
Closest papers by meaning
- DrawingsDreamer: A Unified Multi-View Engineering Drawings Generation Model (NeurIPS 2026 · similarity 0.84)
- Point or Line? Using Line-based Representation for Panoptic Symbol Spotting in CAD Drawings (NeurIPS 2025 · similarity 0.84)
- Revisiting CAD Model Generation by Learning Raster Sketch (AAAI 2025 · similarity 0.84)
- SketchINR: A First Look into Sketches as Implicit Neural Representations (CVPR 2024 · similarity 0.84)
- LineArt: A Knowledge-guided Training-free High-quality Appearance Transfer for Design Drawing with Diffusion Model (CVPR 2025 · similarity 0.82)
- ArchCAD-400K: A Large-Scale CAD drawings Dataset and New Baseline for Panoptic Symbol Spotting (NeurIPS 2025 · similarity 0.82)
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.
