Research map / Image generation / Text-to-image diffusion
Perceptual Pyramid Adversarial Networks for Text-to-Image Synthesis
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 |
|---|---|---|---|
| add Cumulative Blocks (ablation) | PPAN · CUB 256x256 | Inception Score: 4.00 -> 4.03 | Helped |
| add Cumulative Blocks (ablation) | PPAN · CUB 64x64 | Inception Score: 3.27 -> 3.62 | Helped |
| add Cumulative Blocks (ablation) | PPAN · CUB 256x256 | Visual-semantic Similarity: 0.239 -> 0.257 | Helped |
| add perceptual loss L4 (ablation) | PPAN · CUB 256x256 | Inception Score: 4.03 -> 4.15 | Helped |
| add perceptual loss L4 (ablation) | PPAN · CUB 256x256 | Visual-semantic Similarity: 0.257 -> 0.283 | Helped |
| add perceptual loss L4 (ablation) | PPAN · CUB 128x128 | Inception Score: 4.00 -> 4.27 | Helped |
| add class information loss L3 for D3 (ablation) | PPAN · CUB 256x256 | Inception Score: 4.15 -> 4.38 | Helped |
| add class information loss L3 for D3 (ablation) | PPAN · CUB 64x64 | Inception Score: 3.80 -> 3.74 | Mixed |
| add L4 and L3 on every output resolution (ablation) | PPAN · CUB | Inception Score | No clear effect |
| use PPAN single generator instead of HDGAN | PPAN · CUB test | Inception Score: 4.15 -> 4.38 | Helped |
Closest papers by meaning
- Photographic Text-to-Image Synthesis With a Hierarchically-Nested Adversarial Network (CVPR 2018 · similarity 0.90)
- DF-GAN: A Simple and Effective Baseline for Text-to-Image Synthesis (CVPR 2022 · similarity 0.90)
- Adversarial Learning of Semantic Relevance in Text to Image Synthesis (AAAI 2019 · similarity 0.90)
- You Only Need Adversarial Supervision for Semantic Image Synthesis (ICLR 2021 · similarity 0.89)
- Adversarial Text to Continuous Image Generation (CVPR 2024 · similarity 0.87)
- Unlocking Pre-trained Image Backbones for Semantic Image Synthesis (CVPR 2024 · 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.