Research map / Image generation / Autoregressive and latent generation
Semantically Multi-Modal Image 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 progressively decreasing group counts in the decoder | GroupDNet · DeepFashion | FID: 9.84 -> 9.50 | Helped |
| Use a group encoder with class-specific latent codes | GroupDNet · DeepFashion | mOCD: 0.0065 -> 0.0033 | Helped |
| Use GroupDNet rather than SPADE | label-to-image synthesis · Cityscapes | FID: 58.10 -> 49.81 | Helped |
| Use GroupDNet rather than SPADE | label-to-image synthesis · ADE20K | FID: 33.49 -> 42.17 | Hurt |
| Add splitting of the input image into semantic-class images (ablation) | GroupDNet encoder · DeepFashion | mCSD: 0.0054 -> 0.0264 | Helped |
| Add splitting of the input image into semantic-class images (ablation) | GroupDNet encoder · DeepFashion | mOCD: 0.0189 -> 0.0033 | Helped |
| Produce mean and variance maps instead of pooled vectors (ablation) | GroupDNet encoder · DeepFashion | FID: 11.01 -> 9.50 | Helped |
| Replace normalization layers with group normalization (ablation) | GroupDNet · DeepFashion | FID: 9.50 -> 10.33 | Hurt |
| Remove synchronized batch normalization (ablation) | GroupDNet · DeepFashion | FID: 9.50 -> 9.76 | Hurt |
| Remove spectral normalization (ablation) | GroupDNet · DeepFashion | mOCD: 0.0033 -> 0.0153 | Hurt |
| Remove spectral normalization (ablation) | GroupDNet · DeepFashion | mCSD: 0.0264 -> 0.0290 | Mixed |
| Remove spectral normalization (ablation) | GroupDNet · DeepFashion | SHE: 81.2 -> 46.3 | Hurt |
Closest papers by meaning
- Diverse Semantic Image Synthesis via Probability Distribution Modeling (CVPR 2021 · similarity 0.86)
- Unlocking Pre-trained Image Backbones for Semantic Image Synthesis (CVPR 2024 · similarity 0.85)
- Prompting Multi-Modal Image Segmentation with Semantic Grouping (AAAI 2024 · similarity 0.84)
- Seg4Diff: Unveiling Open-Vocabulary Semantic Segmentation in Text-to-Image Diffusion Transformers (NeurIPS 2025 · similarity 0.83)
- Exploring the Deep Fusion of Large Language Models and Diffusion Transformers for Text-to-Image Synthesis (CVPR 2025 · similarity 0.83)
- Variational Partial Group Convolutions for Input-Aware Partial Equivariance of Rotations and Color-Shifts (ICML 2024 · similarity 0.83)
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.