Research map / Image generation / Editing and style transfer
RCEdit-500K: Reference Completion for Image-Conditioned Image Editing
Jingxu Zhang, Daneul Kim, Yueming Pan, DONG CHEN, Kai Qiu, Yang Liu, Yifan Yang, Qi Dai, Xiaoyan Sun, Chong Luo
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 |
|---|---|---|---|
| fine-tune with RCEdit-500K via LoRA | Flux-Kontext · MultiBanana two-reference | Avg: 2.82 -> 4.04 | Improved |
| fine-tune with RCEdit-500K via LoRA | Qwen-Image-Edit-2509 · MultiBanana two-reference | Avg: 4.31 -> 4.56 | Improved |
| full fine-tune Janus-Pro on RCEdit-500K | Janus-Pro · MultiBanana two-reference | Avg: 1.03 -> 4.43 | Improved |
| add five-dimensional VLM post-filtering (ablation) | RCEdit-500K Flux-Kontext LoRA · MultiBanana two-reference (Flux-Kontext LoRA) | Avg: 3.62 -> 4.04 | Improved |
| add five-dimensional VLM post-filtering (ablation) | RCEdit-500K Flux-Kontext LoRA · MultiBanana two-reference (Flux-Kontext LoRA) | BSM: 3.62 -> 4.76 | Improved |
| add weak-instruction augmentation (ablation) | RCEdit-500K Flux-Kontext LoRA · MultiBanana two-reference (Flux-Kontext LoRA) | Avg: 3.74 -> 4.04 | Improved |
| add weak-instruction augmentation (ablation) | RCEdit-500K Flux-Kontext LoRA · MultiBanana two-reference (Flux-Kontext LoRA) | IA: 3.67 -> 3.52 | Hurt |
| add weak-instruction augmentation (ablation) | RCEdit-500K Flux-Kontext LoRA · MultiBanana two-reference (Flux-Kontext LoRA) | RC: 3.35 -> 3.42 | Improved |
| train on ImgEdit ICIE subset | Flux-Kontext LoRA · MultiBanana two-reference | Avg: 2.82 -> 2.92 | Improved |
| train on AnyEdit ICIE subset | Flux-Kontext LoRA · MultiBanana two-reference | IA: 2.94 -> 2.17 | Hurt |
| train on AnyEdit ICIE subset | Flux-Kontext LoRA · MultiBanana two-reference | BSM: 2.39 -> 4.55 | Improved |
| train on RCEdit-60K instead of AnyEdit/ImgEdit | Flux-Kontext LoRA · MultiBanana two-reference | Avg: 2.97 -> 3.68 | Improved |
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.
- DreamOmni2 — DreamOmni2: Multimodal Instruction-based Generation and Editing (CVPR 2026) · 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
- ImgEdit: A Unified Image Editing Dataset and Benchmark (NeurIPS 2025 · similarity 0.90)
- InsightEdit: Towards Better Instruction Following for Image Editing (CVPR 2025 · similarity 0.89)
- UltraEdit: Instruction-based Fine-Grained Image Editing at Scale (NeurIPS 2024 · similarity 0.89)
- Multi-Reward as Condition for Instruction-based Image Editing (ICLR 2025 · similarity 0.89)
- AnyEdit: Mastering Unified High-Quality Image Editing for Any Idea (CVPR 2025 · similarity 0.89)
- PairEdit: Learning Semantic Variations for Exemplar-based Image Editing (NeurIPS 2025 · similarity 0.88)
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.
