Research map / Image generation / Editing and style transfer
InstaGAN: Instance-aware Image-to-Image Translation
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 |
|---|---|---|---|
| Replace CycleGAN with InstaGAN using instance masks | CCP image translation · CCP pants→shorts test | classification score: 0.085 -> 0.232 | Helped |
| Replace CycleGAN with InstaGAN using instance masks | COCO image translation · COCO sheep→giraffe test | masked classification score: 0.594 -> 0.781 | Helped |
| Add a set-based instance-aware architecture instead of translating the union of masks (ablation) | InstaGAN image translation · Multi-instance transfiguration | instance separation and shape quality | Helped |
| Add context preserving loss (ablation) | InstaGAN image translation · Multi-instance transfiguration | background preservation | Helped |
| Add sequential mini-batch translation during training and inference (ablation) | InstaGAN image translation · Images with many instances | generalization performance | Helped |
| Add sequential mini-batch translation during training (ablation) | InstaGAN image translation · Images with few instances | translation quality | Helped |
| Use one-step training with sequential inference instead of sequential training and inference (ablation) | InstaGAN · Multi-instance translation | visual translation quality | Hurt |
| Use sequential training with one-step inference instead of sequential training and inference (ablation) | InstaGAN · Images with many instances | generalization performance | Hurt |
| Order instance mini-batches by decreasing spatial size rather than randomly (ablation) | InstaGAN sequential translation · Multi-instance translation | translation performance | Helped |
| Apply spectral normalization to the discriminator (ablation) | InstaGAN discriminator · Image-to-image translation experiments | performance | Helped |
| Apply spectral normalization to the generator (ablation) | InstaGAN generator · Image-to-image translation experiments | performance | No clear effect |
Closest papers by meaning
- DA-GAN: Instance-Level Image Translation by Deep Attention Generative Adversarial Networks (CVPR 2018 · similarity 0.91)
- Self-Supervised Dense Consistency Regularization for Image-to-Image Translation (CVPR 2022 · similarity 0.89)
- TraVeLGAN: Image-To-Image Translation by Transformation Vector Learning (CVPR 2019 · similarity 0.89)
- Unsupervised Image-to-Image Translation With Generative Prior (CVPR 2022 · similarity 0.88)
- Rethinking the Paradigm of Content Constraints in Unpaired Image-to-Image Translation (AAAI 2024 · similarity 0.87)
- Compressing Image-to-Image Translation GANs Using Local Density Structures on Their Learned Manifold (AAAI 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.