Research map / Image generation / Autoregressive and latent generation
Learning to Incorporate Structure Knowledge for Image Inpainting
Read the paperCheck 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 |
|---|---|---|---|
| Use multi-task image and structure generation rather than image-only generation (ablation) | Image generator baseline · Regular-mask test sets | FID: 15.25 -> 13.66 | Helped |
| Use multi-task image and structure generation rather than image-only generation (ablation) | Image generator baseline · Regular-mask test sets | VIF: 0.823 -> 0.834 | Helped |
| Add structure embedding to multi-task generation (ablation) | MT · Regular-mask test sets | FID: 13.66 -> 12.29 | Helped |
| Add structure embedding to multi-task generation (ablation) | MT · Regular-mask test sets | VIF: 0.834 -> 0.844 | Helped |
| Add attention to multi-task generation with structure embedding (ablation) | MT, SE · Regular-mask test sets | FID: 12.29 -> 11.98 | Helped |
| Add attention to multi-task generation with structure embedding (ablation) | MT, SE · Regular-mask test sets | VIF: 0.844 -> 0.848 | Helped |
| Use the full structure-aware method instead of EG | Image inpainting · Places2 irregular | FID: 13.74 -> 4.883 | Helped |
| Use the full structure-aware method instead of EG | Image inpainting · Places2 regular | FID: 7.536 -> 7.423 | Helped |
| Use the full structure-aware method instead of EG | Image inpainting · CelebA irregular | FID: 2.443 -> 1.227 | Helped |
| Use the full structure-aware method instead of EG | Image inpainting · CelebA regular | FID: 2.421 -> 1.654 | Helped |
Closest papers by meaning
- Incremental Transformer Structure Enhanced Image Inpainting With Masking Positional Encoding (CVPR 2022 · similarity 0.88)
- Foreground-Aware Image Inpainting (CVPR 2019 · similarity 0.88)
- SyFormer: Structure-Guided Synergism Transformer for Large-Portion Image Inpainting (AAAI 2024 · similarity 0.85)
- Paint by Inpaint: Learning to Add Image Objects by Removing Them First (CVPR 2025 · similarity 0.85)
- MTADiffusion: Mask Text Alignment Diffusion Model for Object Inpainting (CVPR 2025 · similarity 0.84)
- Reconstructing the Image Stitching Pipeline: Integrating Fusion and Rectangling into a Unified Inpainting Model (NeurIPS 2024 · similarity 0.84)
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.