Research map / Image generation / Autoregressive and latent generation
PasteGAN: A Semi-Parametric Method to Generate Image from Scene Graph
Read the paperCheck 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 PasteGAN rather than sg2im | scene-graph image generation · COCO-Stuff | Inception Score: 6.7 ± 0.1 -> 9.1 ± 0.2 | Helped |
| Use PasteGAN rather than sg2im | scene-graph image generation · Visual Genome | FID: 71.27 -> 58.53 | Helped |
| Provide ground-truth boxes to PasteGAN | PasteGAN · COCO-Stuff | FID: 50.94 -> 38.29 | Helped |
| Add context-aware Crop Selection instead of selecting random same-category crops (ablation) | PasteGAN · COCO-Stuff | Inception Score: 7.1 ± 0.3 -> 9.1 ± 0.2 | Helped |
| Add context-aware Crop Selection instead of selecting random same-category crops (ablation) | PasteGAN · COCO-Stuff | FID: 96.75 -> 50.94 | Helped |
| Use random same-category crops instead of context-aware Crop Selection (ablation) | PasteGAN · COCO-Stuff | Diversity Score | No clear effect |
| Add Object² Refiner to fuse crop features using object relationships (ablation) | PasteGAN · COCO-Stuff | Inception Score: 8.3 ± 0.3 -> 9.1 ± 0.2 | Helped |
| Add Object² Refiner to fuse crop features using object relationships (ablation) | PasteGAN · COCO-Stuff | FID: 61.28 -> 50.94 | Helped |
| Add Object-Image Fuser (ablation) | PasteGAN · COCO-Stuff | Inception Score: 8.7 ± 0.2 -> 9.1 ± 0.2 | Helped |
| Add Object-Image Fuser (ablation) | PasteGAN · COCO-Stuff | FID: 56.14 -> 50.94 | Helped |
Closest papers by meaning
- Scene Graph Disentanglement and Composition for Generalizable Complex Image Generation (NeurIPS 2024 · similarity 0.88)
- The Scene Language: Representing Scenes with Programs, Words, and Embeddings (CVPR 2025 · similarity 0.87)
- Scene Graph-Grounded Image Generation (AAAI 2025 · similarity 0.87)
- From Pixels to Graphs: Open-Vocabulary Scene Graph Generation with Vision-Language Models (CVPR 2024 · similarity 0.87)
- Scene Graph Generation with Role-Playing Large Language Models (NeurIPS 2024 · similarity 0.87)
- GraphDreamer: Compositional 3D Scene Synthesis from Scene Graphs (CVPR 2024 · similarity 0.87)
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.