Research map / Image generation / Text-to-image diffusion
Conditional Image Repainting via Semantic Bridge and Piecewise Value Function
Shuchen Weng, Wenbo Li, Dawei Li, Hongxia Jin, Boxin Shi
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 |
|---|---|---|---|
| add the full SEBE-GAIN-Cls-V3(9) design | best baseline variant · COCO-Stuff | R-precision (%): 61.13 -> 65.43 | Improved |
| add the full method (SEBE, GAIN, Cls, V3(9)) | SEBE w/ DotPrdct · CUB-200-2011 | M-score: 32.86 -> 24.6 | Improved |
| add the full method | CrsMod · COCO-Stuff | M-score: 78.54 -> 67.96 | Improved |
| add SEBE attention to cross-modality attention estimation (ablation) | CrsMod · CUB-200-2011 | M-score: 31.38 -> 24.6 | Mixed |
| add cross-modality attention in SEBE (ablation) | SEBE w/o CrsMod · COCO-Stuff | M-score: 72.11 -> 67.96 | Hurt |
| add attention selection (SAS) (ablation) | SEBE w/o SAS · COCO-Stuff | R-precision (%): 61.13 -> 65.43 | Improved |
| add attention selection (SAS) (ablation) | SEBE w/o SAS · CUB-200-2011 | M-score: 34.18 -> 24.6 | Improved |
| use classification discriminator instead of segmentation discriminator (ablation) | V3(9) · CUB-200-2011 | M-score: 33.53 -> 24.6 | Improved |
| use classification discriminator instead of segmentation discriminator (ablation) | V2(4) · CUB-200-2011 | M-score: 28.1 -> 27.17 | Improved |
| use V3(9) instead of V2(4) value function (ablation) | Cls discriminator · CUB-200-2011 | M-score: 27.17 -> 24.6 | Improved |
| use V3(9) instead of V2(4) value function (ablation) | Seg discriminator · COCO-Stuff | M-score: 75.16 -> 76.35 | No clear effect |
| use V3(9) instead of V'2(5) value function (ablation) | Cls discriminator · COCO-Stuff | M-score: 81.22 -> 67.96 | Improved |
Closest papers by meaning
- Conditional Image Repainting (TPAMI 2023 · similarity 0.88)
- Generative Semantic Manipulation with Mask-Contrasting GAN (ECCV 2018 · similarity 0.87)
- UniCoRN: A Unified Conditional Image Repainting Network (CVPR 2022 · similarity 0.87)
- Towards Diverse and Natural Image Descriptions via a Conditional GAN (ICCV 2017 · similarity 0.87)
- Photographic Text-to-Image Synthesis With a Hierarchically-Nested Adversarial Network (CVPR 2018 · similarity 0.87)
- Spin: Diffusion-based Semantic Image Painting Through Independent Information Injection (AAAI 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.
