Research map / Image generation / GANs
Lightweight Generative Adversarial Networks for Text-Guided Image Manipulation
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 the proposed word-level discriminator instead of removing it (ablation) | Lightweight GAN · CUB | FID: 9.20 -> 8.02 | Helped |
| use the proposed word-level discriminator instead of removing it (ablation) | Lightweight GAN · COCO | FID: 15.60 -> 12.39 | Helped |
| use the proposed word-level discriminator instead of a text-adaptive discriminator (ablation) | Lightweight GAN · CUB | FID: 8.84 -> 8.02 | Helped |
| use the proposed word-level discriminator instead of a text-adaptive discriminator (ablation) | Lightweight GAN · COCO | FID: 14.95 -> 12.39 | Helped |
| use the proposed word-level discriminator instead of the ControlGAN word-level discriminator (ablation) | Lightweight GAN · CUB | FID: 8.72 -> 8.02 | Helped |
| use the proposed word-level discriminator instead of the ControlGAN word-level discriminator (ablation) | Lightweight GAN · COCO | FID: 13.68 -> 12.39 | Helped |
| reduce model size to the reported Ours* configuration (ablation) | Lightweight GAN · CUB | generator parameters: 18.5M -> 5.4M | Helped |
| reduce model size to the reported Ours* configuration (ablation) | Lightweight GAN · COCO | generator parameters: 30.1M -> 7.4M | Helped |
| use the lightweight proposed architecture rather than ManiGAN | Text-guided image manipulation · CUB | FID: 9.75 -> 8.02 | Helped |
| use the lightweight proposed architecture rather than ManiGAN | Text-guided image manipulation · COCO | FID: 25.08 -> 12.39 | Helped |
Closest papers by meaning
- Text-Adaptive Generative Adversarial Networks: Manipulating Images with Natural Language (NeurIPS 2018 · similarity 0.93)
- ManiGAN: Text-Guided Image Manipulation (CVPR 2020 · similarity 0.91)
- Controllable Text-to-Image Generation (NeurIPS 2019 · similarity 0.89)
- DF-GAN: A Simple and Effective Baseline for Text-to-Image Synthesis (CVPR 2022 · similarity 0.88)
- TiGAN: Text-Based Interactive Image Generation and Manipulation (AAAI 2022 · similarity 0.88)
- E$^2$GAN: Efficient Training of Efficient GANs for Image-to-Image Translation (ICML 2024 · similarity 0.85)
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.