Research map / Image generation / Editing and style transfer
Multi-Channel Attention Selection GAN With Cascaded Semantic Guidance for Cross-View Image Translation
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 target semantic map alongside source image as generator input (ablation) | Pix2pix-style generator · Dayton a2g ablation subset | SSIM: 0.4555 -> 0.5374 | Helped |
| add cycled semantic generation (ablation) | Semantic-guided generation · Dayton a2g ablation subset | SSIM: 0.5374 -> 0.5438 | Helped |
| add uncertainty-guided pixel loss (ablation) | Cycled semantic generation · Dayton a2g ablation subset | SSIM: 0.5438 -> 0.5522 | Helped |
| add multi-channel attention selection (ablation) | Uncertainty-guided generation · Dayton a2g ablation subset | SSIM: 0.5522 -> 0.5989 | Helped |
| add total variation regularization (ablation) | Multi-channel attention model · Dayton a2g ablation subset | SSIM: 0.5989 -> 0.6047 | Helped |
| add multi-scale spatial pooling (ablation) | Total-variation-regularized model · Dayton a2g ablation subset | SSIM: 0.6047 -> 0.6167 | Helped |
| increase attention channels from 0 to 1 (ablation) | SelectionGAN · Dayton a2g ablation subset | SSIM: 0.5438 -> 0.5522 | Helped |
| increase attention channels from 1 to 5 (ablation) | SelectionGAN · Dayton a2g ablation subset | SSIM: 0.5522 -> 0.5901 | Helped |
| increase attention channels from 5 to 10 (ablation) | SelectionGAN · Dayton a2g ablation subset | SSIM: 0.5901 -> 0.5986 | Helped |
| increase attention channels from 10 to 32 (ablation) | SelectionGAN · Dayton a2g ablation subset | SSIM: 0.5986 -> 0.5950 | Hurt |
| use the stage-II output rather than only the stage-I output (ablation) | SelectionGAN with multi-channel attention selection · Dayton a2g ablation subset | SSIM: 0.5551 -> 0.5989 | Helped |
| use the stage-II output rather than only the stage-I output (ablation) | SelectionGAN with total variation regularization · Dayton a2g ablation subset | SSIM: 0.5680 -> 0.6047 | Helped |
Closest papers by meaning
- Edge Guided GANs with Contrastive Learning for Semantic Image Synthesis (ICLR 2023 · similarity 0.87)
- Cross-View Image Synthesis Using Conditional GANs (CVPR 2018 · similarity 0.87)
- TraVeLGAN: Image-To-Image Translation by Transformation Vector Learning (CVPR 2019 · similarity 0.86)
- Cross-Modal Contrastive Learning for Text-to-Image Generation (CVPR 2021 · similarity 0.86)
- Saliency-Guided Image Translation (CVPR 2021 · similarity 0.85)
- SynthRGB-T: Language-Vision Guided Image Translation for Diversity Synthesis (CVPR 2026 · 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.