Research map / Image generation / Editing and style transfer
SCAM! Transferring Humans between Images with Semantic Cross Attention Modulation
Nicolas Dufour, David Picard, Vicky Kalogeiton
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 convolutions to the SAT encoder (ablation) | SCAM · iDesigner | R-FID: 26.3 -> 13.2 | Improved |
| add convolutions to the SAT encoder (ablation) | SCAM · CelebAMask-HQ | R-FID: 15.5 -> 15.7 | No clear effect |
| add self-attention to the SAT encoder (ablation) | SCAM · iDesigner | R-FID: 12.7 -> 13.2 | No clear effect |
| add self-attention to the SAT encoder (ablation) | SCAM · CelebAMask-HQ | R-FID: 16.5 -> 15.7 | Improved |
| add SAT block inside the SCAM generator (ablation) | SCAM · iDesigner | R-FID: 15.6 -> 13.2 | Improved |
| add SAT block inside the SCAM generator (ablation) | SCAM · iDesigner | S-FID: 27.7 -> 27.1 | Improved |
| use SAT encoder instead of SEAN encoder (ablation) | SCAM · iDesigner | R-FID: 27.6 -> 13.2 | Improved |
| add L1 loss (ablation) | SCAM · iDesigner | R-FID: 16.1 -> 13.2 | Improved |
| add perceptual loss (ablation) | SCAM · iDesigner | S-FID: 91.3 -> 27.1 | Improved |
| add perceptual loss (ablation) | SCAM · iDesigner | PSNR: 22.9 -> 21.0 | Hurt |
| increase latents per label k from 4 to 32 (ablation) | SCAM · iDesigner | R-FID: 15.9 -> 9.7 | Improved |
| increase latents per label k from 4 to 32 (ablation) | SCAM · iDesigner | S-FID: 27.3 -> 30.3 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SEAN — SEAN: Image Synthesis With Semantic Region-Adaptive Normalization (CVPR 2020) · better on 10 of 10 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Cross-view Masked Diffusion Transformers for Person Image Synthesis (ICML 2024 · similarity 0.86)
- Cross-Attention Head Position Patterns Can Align with Human Visual Concepts in Text-to-Image Generative Models (ICLR 2025 · similarity 0.86)
- Coarse-to-Fine Latent Diffusion for Pose-Guided Person Image Synthesis (CVPR 2024 · similarity 0.85)
- Self-Cross Diffusion Guidance for Text-to-Image Synthesis of Similar Subjects (CVPR 2025 · similarity 0.85)
- IMAGPose: A Unified Conditional Framework for Pose-Guided Person Generation (NeurIPS 2024 · similarity 0.85)
- Instruct-Imagen: Image Generation with Multi-modal Instruction (CVPR 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.
