Research map / Image generation / Autoregressive and latent generation
High-Resolution Image Inpainting Using Multi-Scale Neural Patch Synthesis
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 |
|---|---|---|---|
| Use joint content and neural-patch texture optimization | Our Method · Paris StreetView | Mean L1 Loss: 10.33% -> 10.01% | Helped |
| Use joint content and neural-patch texture optimization | Our Method · Paris StreetView | Mean L2 Loss: 2.35% -> 2.21% | Helped |
| Use joint content and neural-patch texture optimization | Our Method · Paris StreetView | PSNR: 17.59 dB -> 18.00 dB | Helped |
| Add the global content constraint (ablation) | Our Method · Image inpainting | visual structural correctness | Helped |
| Increase the content constraint relative to the texture constraint (ablation) | Our Method · Image inpainting | high-frequency detail | Hurt |
| Increase the content constraint relative to the texture constraint (ablation) | Our Method · Image inpainting | consistency with initial content prediction | Helped |
| Increase the texture constraint relative to the content constraint (ablation) | Our Method · Image inpainting | sharpness | Helped |
| Increase the texture constraint relative to the content constraint (ablation) | Our Method · Image inpainting | structural correctness | Hurt |
| Add adversarial loss to content-network training (ablation) | Our Method · Image inpainting | final visual quality | Helped |
| Use pre-trained VGG-19 instead of the content network as the texture network (ablation) | Our Method · Image inpainting | texture reconstruction quality | Helped |
| Combine relu3_1 and relu4_1 instead of using a single feature layer (ablation) | Our Method · Image inpainting | texture reconstruction accuracy | Helped |
Closest papers by meaning
- Generative Image Inpainting With Contextual Attention (CVPR 2018 · similarity 0.92)
- Learning Pyramid-Context Encoder Network for High-Quality Image Inpainting (CVPR 2019 · similarity 0.89)
- Contextual Residual Aggregation for Ultra High-Resolution Image Inpainting (CVPR 2020 · similarity 0.89)
- SyFormer: Structure-Guided Synergism Transformer for Large-Portion Image Inpainting (AAAI 2024 · similarity 0.86)
- Reconstructing the Image Stitching Pipeline: Integrating Fusion and Rectangling into a Unified Inpainting Model (NeurIPS 2024 · similarity 0.86)
- InpDiffusion: Image Inpainting Localization via Conditional Diffusion Models (AAAI 2025 · similarity 0.86)
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.