Research map / Image generation / GANs
Variational Hetero-Encoder Randomized GANs for Joint Image-Text Modeling
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 hierarchical-semantic raster-scan generation instead of VHE-StackGAN++ (ablation) | VHE-GAN · Flower, CUB, COCO | IS: 3.56 -> 3.72; 4.20 -> 4.41; 12.63 -> 27.16 | Helped |
| Use hierarchical-semantic raster-scan generation instead of VHE-StackGAN++ (ablation) | VHE-GAN · Flower, CUB, COCO | FID: 38.66 -> 35.13; 12.93 -> 12.02; 78.02 -> 75.88 | Helped |
| Add stacking structure to raster-scan generation (ablation) | VHE-simple-raster-scan-GAN · Flower, CUB, COCO | IS: 3.62 -> 3.72; 4.31 -> 4.41; 20.13 -> 27.16 | Helped |
| Add stacking structure to raster-scan generation (ablation) | VHE-simple-raster-scan-GAN · Flower, CUB, COCO | FID: 36.18 -> 35.13; 12.35 -> 12.02; 77.18 -> 75.88 | Helped |
| Replace stacked image generator with vanilla GAN (ablation) | VHE-GAN · Flower, CUB, COCO | IS: 3.01 -> 3.56; 3.52 -> 4.20; 6.36 -> 12.63 | Helped |
| Replace stacked image generator with vanilla GAN (ablation) | VHE-GAN · Flower, CUB, COCO | FID: 52.15 -> 38.66; 21.24 -> 12.93; 97.15 -> 78.02 | Helped |
| Replace RNN text features with PGBN features in separate training (ablation) | StackGAN++ · Flower, CUB, COCO | IS: 3.26 -> 3.29; 3.84 -> 3.92; 8.30 -> 10.63 | Helped |
| Replace RNN text features with PGBN features in separate training (ablation) | StackGAN++ · Flower, CUB, COCO | FID: 48.68 -> 41.04; 15.30 -> 13.79; 81.59 -> 79.65 | Mixed |
| Jointly train text feature extractor, image encoder, discriminator, and generator (ablation) | PGBN+StackGAN++ · Flower, CUB, COCO | IS: 3.29 -> 3.56; 3.92 -> 4.20; 10.63 -> 12.63 | Helped |
| Jointly train text feature extractor, image encoder, discriminator, and generator (ablation) | PGBN+StackGAN++ · Flower, CUB, COCO | FID: 41.04 -> 38.66; 13.79 -> 12.93; 79.65 -> 78.02 | Helped |
Closest papers by meaning
- DF-GAN: A Simple and Effective Baseline for Text-to-Image Synthesis (CVPR 2022 · similarity 0.89)
- Generating Diverse High-Fidelity Images with VQ-VAE-2 (NeurIPS 2019 · similarity 0.87)
- E$^2$GAN: Efficient Training of Efficient GANs for Image-to-Image Translation (ICML 2024 · similarity 0.87)
- StyleGAN-T: Unlocking the Power of GANs for Fast Large-Scale Text-to-Image Synthesis (ICML 2023 · similarity 0.87)
- Improving GAN Training with Probability Ratio Clipping and Sample Reweighting (NeurIPS 2020 · similarity 0.87)
- Adversarial Text to Continuous Image Generation (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.