Research map / Image generation / Autoregressive and latent generation
Image Inpainting Guided by Coherence Priors of Semantics and Textures
Liang Liao, Jing Xiao, Zheng Wang, Chia-Wen Lin, Shin'ichi Satoh
Check 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 SWAP to the joint multi-scale base method (ablation) | Image inpainting on Outdoor Scenes · Outdoor Scenes | PSNR: 17.43 -> 20.58 | Improved |
| Add SWAP to the joint multi-scale base method (ablation) | Image inpainting on Outdoor Scenes · Outdoor Scenes | FID: 57.31 -> 39.46 | Improved |
| Replace attention across all semantics with semantic-wise SWAP (ablation) | Image inpainting on Outdoor Scenes · Outdoor Scenes | PSNR: 19.77 -> 20.58 | Improved |
| Replace attention across all semantics with semantic-wise SWAP (ablation) | Image inpainting on Outdoor Scenes · Outdoor Scenes | FID: 43.54 -> 39.46 | Improved |
| Add coherence loss to the SWAP variant (ablation) | Image inpainting on Outdoor Scenes · Outdoor Scenes | PSNR: 20.58 -> 21.18 | Improved |
| Add coherence loss to the SWAP variant (ablation) | Image inpainting on Outdoor Scenes · Outdoor Scenes | FID: 39.46 -> 38.15 | Improved |
| Use semantic-wise attention and coherence losses rather than the base framework (ablation) | Image inpainting on Outdoor Scenes · Outdoor Scenes | SSIM: 0.65 -> 0.81 | Improved |
| Use the full coherence-prior method rather than SGE-Net | Image inpainting on Cityscapes, centering holes · Cityscapes, centering holes | SSIM: 0.85 -> 0.84 | Hurt |
| Use the full coherence-prior method rather than SGE-Net | Image inpainting on Outdoor Scenes, centering holes · Outdoor Scenes, centering holes | SSIM: 0.81 -> 0.81 | No clear effect |
| Jointly optimize inpainting and segmentation with the coherence-prior method | Semantic segmentation · Outdoor Scenes | mIoU%: 0.68 -> 0.71 | Improved |
| Jointly optimize inpainting and segmentation with the coherence-prior method | Semantic segmentation · Cityscapes | mIoU%: 0.53 -> 0.57 | Improved |
Closest papers by meaning
- Guidance and Evaluation: Semantic-Aware Image Inpainting for Mixed Scenes (ECCV 2020 · similarity 0.91)
- Coherent Semantic Attention for Image Inpainting (ICCV 2019 · similarity 0.88)
- Contextual-based Image Inpainting: Infer, Match, and Translate (ECCV 2018 · similarity 0.88)
- Semantic Image Inpainting With Deep Generative Models (CVPR 2017 · similarity 0.87)
- Structure Matters: Tackling the Semantic Discrepancy in Diffusion Models for Image Inpainting (CVPR 2024 · similarity 0.85)
- Visual Jenga: Discovering Object Dependencies via Counterfactual Inpainting (NeurIPS 2025 · 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.
