Research map / Image generation / Editing and style transfer
Semantic Scale Space: A Framework for Controllable Image Abstraction
Kazu Mishiba
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 |
|---|---|---|---|
| use AGSS donor-gated fine-to-coarse smoothing instead of canonical baselines at strong level | AGSS · SBD (strong, RHI-matched) | BPF-ODS: 0.545 (WLS) -> 0.594 | Improved |
| same, medium level | AGSS · SBD (medium, RHI-matched) | BPF-ODS: 0.541 (WLS) -> 0.575 | Improved |
| same, weak level | AGSS · SBD (weak, RHI-matched) | BPF-ODS: 0.529 (WLS) -> 0.535 | No clear effect |
| lower geometric drift with AGSS at strong level | AGSS · SBD (strong, RHI-matched) | Drift (px): 85.65 (WLS) -> 82.98 | Improved |
| increase abstraction strength from weak to strong | AGSS · SBD | BPF-ODS: 0.535 -> 0.594 | Improved |
| increase abstraction strength from weak to strong | PM · SBD | BPF-ODS: 0.529 -> 0.526 | No clear effect |
| swap donor-gated averaging for Domain Transform with the same MuGE cues (ablation) | DT+MuGE vs AGSS · SBD (strong) | BPF-ODS: 0.555 (DT+MuGE) -> 0.594 (AGSS) | Improved |
| DT+MuGE at weak level vs canonical baselines | DT+MuGE · SBD (weak) | BPF-ODS: 0.529 -> 0.497 | Hurt |
| apply AGSS pre-abstraction before NPR stylizers | XDoG, white-box cartoonization, sketching · DIV2K 2AFC | preference rate (%): 72.9% preferred | Improved |
Closest papers by meaning
- Layered Rendering Diffusion Model for Controllable Zero-Shot Image Synthesis (ECCV 2024 · similarity 0.88)
- Semantics Lead the Way: Harmonizing Semantic and Texture Modeling with Asynchronous Latent Diffusion (CVPR 2026 · similarity 0.86)
- Move Anything with Layered Scene Diffusion (CVPR 2024 · similarity 0.85)
- MultiDiff: Consistent Novel View Synthesis from a Single Image (CVPR 2024 · similarity 0.85)
- Spin: Diffusion-based Semantic Image Painting Through Independent Information Injection (AAAI 2025 · similarity 0.85)
- Image Inpainting via Tractable Steering of Diffusion Models (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.
