Research map / Image generation / Autoregressive and latent generation
Scribbler: Controlling Deep Image Synthesis With Sketch and Color
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 adversarial loss to content loss (ablation) | Scribbler sketch-to-photo synthesis · Faces, cars, and bedrooms | visual realism and diversity | Helped |
| Use the deeper encoder-decoder architecture instead of Sketch Inversion | Scribbler sketch-to-photo synthesis · Complex bedroom scenes and simplified sketches | sharpness and hallucination of missing details | Helped |
| Add multiple sketch styles to training and fine-tune (ablation) | Scribbler face sketch-to-photo synthesis · Arbitrary hand-drawn face sketches | generalization | Helped |
| Add random cropping and sketch-line cut-off augmentation (ablation) | Scribbler sketch-to-photo synthesis · Sketches with missing details | hallucination of missing details | Helped |
| Add sparse color strokes to the input | Scribbler guided sketch colorization · Bedroom and car sketches | satisfaction of user color constraints | Helped |
| Use adversarial loss during sketch reconstruction (ablation) | Scribbler · Exaggerated or cartoonish face sketches | faithfulness to sketch constraints | Hurt |
| Use adversarial loss to satisfy rare color controls (ablation) | Scribbler colorization · Rare color requests, such as purple cars or red trees | faithfulness to color controls | Hurt |
| Use a weak discriminator and do not condition it on the input | Scribbler adversarial training · Training stability | stability | Helped |
Closest papers by meaning
- TextureGAN: Controlling Deep Image Synthesis With Texture Patches (CVPR 2018 · similarity 0.90)
- Self-Supervised Sketch-to-Image Synthesis (AAAI 2021 · similarity 0.88)
- High-Fidelity Guided Image Synthesis With Latent Diffusion Models (CVPR 2023 · similarity 0.88)
- Picture That Sketch: Photorealistic Image Generation From Abstract Sketches (CVPR 2023 · similarity 0.87)
- Sketch Down the FLOPs: Towards Efficient Networks for Human Sketch (CVPR 2025 · similarity 0.87)
- It's All About Your Sketch: Democratising Sketch Control in Diffusion Models (CVPR 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.