Research map / Image generation / Editing and style transfer
ManiFest: Manifold Deformation for Few-Shot Image Translation
Fabio Pizzati, Jean-François Lalonde, Raoul de Charette
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 |
|---|---|---|---|
| replace MUNIT few-shot training with ManiFest (general) | MUNIT backbone · ACDC Day->Night (|T|=25) | FID: 91.61 -> 81.01 | Improved |
| replace MUNIT few-shot training with ManiFest (exemplar) | MUNIT backbone · ACDC Day->Night (|T|=25) | FID: 128.73 -> 80.57 | Improved |
| ManiFest exemplar vs MUNIT trained on all 400 images | ManiFest · ACDC Day->Night | FID: 87.71 -> 80.57 | Improved |
| ManiFest general vs MUNIT trained on all 400 images | ManiFest · ACDC Day->Night | FID: 79.20 -> 81.01 | No clear effect |
| ManiFest vs FUNIT on Clear->Fog general | ManiFest · ACDC Clear->Fog | FID: 152.90 -> 89.57 | Improved |
| use multiple anchor domains (All) instead of Night anchor, Day->Twilight (ablation) | ManiFest · DZ Day->Twilight | FID: 63.15 -> 60.98 | Improved |
| use multiple anchors for Day->Night (ablation) | ManiFest · ACDC Day->Night | FID: 81.01 -> 83.71 | No clear effect |
| use Rain anchor instead of Night for Day->Night (ablation) | ManiFest · ACDC Day->Night | FID: 81.01 -> 81.38 | No clear effect |
| reduce few-shot set from 25 to 5 images (ablation) | ManiFest · ACDC Day->Night general | FID: 82.95 -> 83.11 | No clear effect |
| reduce few-shot set to a single image (ablation) | ManiFest · ACDC Day->Night general | FID: 82.95 -> 114.5 | Hurt |
| train segmentation on ManiFest-translated night Cityscapes vs MUNIT | HRNet · ACDC-night val | mIoU: +3.09 mIoU | Improved |
| apply GERM residual on synthetic anchor images | ManiFest · ACDC-Night | FID: 142 -> 130 | Improved |
Closest papers by meaning
- Few-Shot Unsupervised Image-to-Image Translation (ICCV 2019 · similarity 0.89)
- COCO-FUNIT: Few-Shot Unsupervised Image Translation with a Content Conditioned Style Encoder (ECCV 2020 · similarity 0.88)
- TransGaGa: Geometry-Aware Unsupervised Image-To-Image Translation (CVPR 2019 · similarity 0.87)
- Preserve and Sculpt: Manifold-Aligned Fine-tuning of Vision-Language Models for Few-Shot Learning (ICLR 2026 · similarity 0.87)
- Memory-Guided Unsupervised Image-to-Image Translation (CVPR 2021 · similarity 0.86)
- TraVeLGAN: Image-To-Image Translation by Transformation Vector Learning (CVPR 2019 · 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.
