Research map / Image generation / GANs
TopoGAN: A Topology-Aware Generative Adversarial Network
Fan Wang, Huidong Liu, Dimitris Samaras, Chao Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 topological GAN loss | WGAN-GP · CREMI generation | Betti score: 0.236 -> 0.015 | Improved |
| add topological GAN loss | WGAN-GP · ISBI12 generation | unbiased MMD: 0.558 -> 0.405 | Improved |
| add topological GAN loss | WGAN-GP · Facade generation | unbiased MMD: 0.137 -> 0.080 | Improved |
| add topological GAN loss | WGAN-GP · Maps generation | Betti score: 0.223 -> 0.177 | Improved |
| add topological GAN loss | WGAN-GP · ISBI12 generation | FID: 83.90 -> 31.90 | Improved |
| add topological GAN loss | WGAN-GP · CREMI generation | FID: 21.64 -> 20.96 | No clear effect |
| add topological GAN loss | WGAN-GP · Retina generation | FID: 179.69 -> 169.21 | No clear effect |
| add topological GAN loss | WGAN-GP · Retina generation | unbiased MMD: 1.735 -> 1.602 | No clear effect |
| train segmenter on TopoGAN synthetic data instead of WGAN-GP synthetic data | U-Net · CREMI segmentation | Dice score: 0.820 -> 0.851 | Improved |
| train on synthetic only instead of real data | U-Net with TopoGAN data · CREMI segmentation | Dice score: 0.896 -> 0.851 | Hurt |
| train on synthetic only instead of real data | U-Net with TopoGAN data · ISBI12 segmentation | Dice score: 0.932 -> 0.933 | No clear effect |
| augment real data with TopoGAN synthetic data | U-Net · Retina segmentation | Dice score: 0.883 -> 0.906 | Improved |
Closest papers by meaning
- Topology Distance: A Topology-Based Approach for Evaluating Generative Adversarial Networks (AAAI 2021 · similarity 0.89)
- A Geometric Analysis of Deep Generative Image Models and Its Applications (ICLR 2021 · similarity 0.89)
- Linear Semantics in Generative Adversarial Networks (CVPR 2021 · similarity 0.87)
- On the "steerability" of generative adversarial networks (ICLR 2020 · similarity 0.87)
- Unlocking Pre-trained Image Backbones for Semantic Image Synthesis (CVPR 2024 · similarity 0.87)
- Conditional Generative Adversarial Network for Structured Domain Adaptation (CVPR 2018 · 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.
