Research map / Image generation / Editing and style transfer
AM-Adapter: Appearance Matching Adapter for Exemplar-based Semantic Image Synthesis in-the-Wild
Siyoon Jin, Jisu Nam, Jiyoung Kim, Dahyun Chung, Yeong-Seok Kim, Joonhyung Park, Heonjeong Chu, Seungryong Kim
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 augmented self-attention (ablation) | ControlNeXt · BDD100K (segmentation-to-image) | I_CLIP: 0.731 -> 0.815 | Improved |
| add augmented self-attention (ablation) | ControlNeXt · BDD100K (segmentation-to-image) | FID: 94.48 -> 78.13 | Improved |
| add joint fine-tuning of augmented self-attention (ablation) | augmented self-attention · BDD100K (segmentation-to-image) | I_CLIP: 0.815 -> 0.808 | Hurt |
| add joint fine-tuning of augmented self-attention (ablation) | augmented self-attention · BDD100K (segmentation-to-image) | FID: 78.13 -> 77.75 | No clear effect |
| add categorical matching cost (ablation) | augmented self-attention · BDD100K (segmentation-to-image) | FID: 78.13 -> 76.76 | Improved |
| add 4D-aggregated matching adapter trained in separate stage (ablation) | augmented self-attention + categorical matching cost · BDD100K (segmentation-to-image) | FID: 76.76 -> 75.89 | Improved |
| add 4D-aggregated matching adapter trained in separate stage (ablation) | augmented self-attention + categorical matching cost · BDD100K (segmentation-to-image) | Self-sim.: 0.043 -> 0.041 | Improved |
| replace key-value replacement (MasaCtrl) with augmented self-attention (ablation) | ControlNeXt · BDD100K (segmentation-to-image) | I_CLIP: 0.872 -> 0.815 | Hurt |
| add DreamMatcher (ablation) | ControlNeXt · BDD100K (segmentation-to-image) | I_CLIP: 0.731 -> 0.724 | No clear effect |
| add exemplar retrieval at inference (ablation) | AM-Adapter · BDD100K (segmentation-to-image) | I_CLIP: 0.741 -> 0.814 | Improved |
| add matching cost guidance at inference (ablation) | AM-Adapter + retrieval · BDD100K (segmentation-to-image) | FID: 77.15 -> 75.89 | Improved |
| add exemplar retrieval at inference (ablation) | AM-Adapter · BDD100K (segmentation-to-image) | FID: 79.05 -> 77.15 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- Ctrl-X — Ctrl-X: Controlling Structure and Appearance for Text-To-Image Generation Without Guidance (NeurIPS 2024) · better on 5 of 5 shared results
- FreeControl — FreeControl: Training-Free Spatial Control of Any Text-to-Image Diffusion Model with Any Condition (CVPR 2024) · better on 1 of 1 shared result
- ControlNet — ControlNet++: Improving Conditional Controls with Efficient Consistency Feedback (ECCV 2024) · better on 1 of 1 shared result
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-Domain Correspondence Learning for Exemplar-Based Image Translation (CVPR 2020 · similarity 0.89)
- Masked and Adaptive Transformer for Exemplar Based Image Translation (CVPR 2023 · similarity 0.87)
- Marginal Contrastive Correspondence for Guided Image Generation (CVPR 2022 · similarity 0.87)
- Novel Object Synthesis via Adaptive Text-Image Harmony (NeurIPS 2024 · similarity 0.85)
- Object-Aware Inversion and Reassembly for Image Editing (ICLR 2024 · similarity 0.85)
- Hummingbird: High Fidelity Image Generation via Multimodal Context Alignment (ICLR 2025 · 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.
