Research map / Image generation / Text-to-image diffusion
Match-and-Fuse: Consistent Generation from Unstructured Image Sets
Kate Feingold, Omri Kaduri, Tali Dekel
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 Feature Guidance (ablation) | Match-and-Fuse · Set consistency benchmark | DINO-MatchSim: 0.76 -> 0.80 | Improved |
| add Multiview Feature Fusion (ablation) | Match-and-Fuse · Set consistency benchmark | DINO-MatchSim: 0.78 -> 0.80 | Improved |
| add the pairwise consistency graph (ablation) | Match-and-Fuse · Set consistency benchmark | DINO-MatchSim: 0.75 -> 0.80 | Improved |
| add the pairwise consistency graph (ablation) | Match-and-Fuse · Set consistency benchmark | DreamSim: 0.82 -> 0.85 | Improved |
| add each ablated component on prompt adherence (ablation) | Match-and-Fuse · Set consistency benchmark | CLIP score: 0.66 -> 0.66 | No clear effect |
| limit graph to 4 random neighbors per node instead of fully connected graph (ablation) | Match-and-Fuse · Consistency vs number of images | DINO-MatchSim | Hurt |
| use Match-and-Fuse instead of Edicho | consistent set editing · Set consistency benchmark | DINO-MatchSim: 0.72 -> 0.80 | Improved |
| use DINO-MatchSim instead of DreamSim as consistency metric | evaluation · Agreement with human votes | agreement %: 84.3 -> 91.4 | Improved |
| add Feature Guidance when matches are sparsified (ablation) | Match-and-Fuse · Match sparsity analysis | DINO-MatchSim | Improved |
Closest papers by meaning
- UniReal: Universal Image Generation and Editing via Learning Real-world Dynamics (CVPR 2025 · similarity 0.86)
- Contextual Flow Matching: Adaptive Step Selection in Flow Models for Efficient Visual Generation (NeurIPS 2026 · similarity 0.85)
- HarmonyView: Harmonizing Consistency and Diversity in One-Image-to-3D (CVPR 2024 · similarity 0.85)
- SpotActor: Training-Free Layout-Controlled Consistent Image Generation (AAAI 2025 · similarity 0.85)
- ViewDiff: 3D-Consistent Image Generation with Text-to-Image Models (CVPR 2024 · similarity 0.85)
- RefDrop: Controllable Consistency in Image or Video Generation via Reference Feature Guidance (NeurIPS 2024 · similarity 0.84)
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.
