Research map / Image generation / Editing and style transfer
Diverse Image-to-Image Translation via Disentangled Representations
Hsin-Ying Lee, Hung-Yu Tseng, Jia-Bin Huang, Maneesh Singh, Ming-Hsuan Yang
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 content discriminator (ablation) | DRIT · Yosemite winter->summer (LPIPS diversity) | LPIPS diversity: .410 -> .424 | Improved |
| add content discriminator (ablation) | DRIT · Edge-to-shoes (reconstruction) | reconstruction error (L1): 0.2076 -> 0.1347 | Improved |
| generate 3x instead of 1x target images with diverse outputs | DRIT domain adaptation · MNIST to MNIST-M | accuracy (%): 86.93 -> 90.21 | Improved |
| generate 5x instead of 3x target images | DRIT domain adaptation · MNIST to MNIST-M | accuracy (%): 90.21 -> 91.54 | Improved |
| generate 5x instead of 1x target images | DRIT domain adaptation · LineMod | mean angle error (deg): 42.06 -> 34.4 | Improved |
| generate 5x instead of 1x target images | DRIT domain adaptation · LineMod | classification accuracy (%): 95.91 -> 98.12 | Improved |
| use DRIT diverse translation instead of CycleGAN for data generation | domain adaptation · MNIST to MNIST-M | accuracy (%): 74.5 -> 86.93 | Improved |
| use paired data (BicycleGAN) instead of unpaired DRIT | encoder disentanglement · Edge-to-shoes | reconstruction error (L1): 0.1347 -> 0.0945 | Improved |
Closest papers by meaning
- DRIT++: Diverse Image-to-Image Translation via Disentangled Representations (IJCV 2020 · similarity 0.97)
- Multi-mapping Image-to-Image Translation via Learning Disentanglement (NeurIPS 2019 · similarity 0.91)
- Image-to-image translation for cross-domain disentanglement (NeurIPS 2018 · similarity 0.90)
- Scaling-Up Disentanglement for Image Translation (ICCV 2021 · similarity 0.90)
- SynthRGB-T: Language-Vision Guided Image Translation for Diversity Synthesis (CVPR 2026 · similarity 0.87)
- Transferring disentangled representations: bridging the gap between synthetic and real images (NeurIPS 2024 · similarity 0.87)
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.
