Research map / Image generation / Text-to-image diffusion
ELITE: Encoding Visual Concepts into Textual Embeddings for Customized Text-to-Image Generation
Yuxiang Wei, Yabo Zhang, Zhilong Ji, Jinfeng Bai, Lei Zhang, Wangmeng Zuo
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 |
|---|---|---|---|
| learn multiple words from multiple CLIP layers instead of from the single deepest layer, and use primary word only (ablation) | ELITE global mapping · DreamBench | CLIP-T: 0.212 -> 0.257 | Improved |
| learn multiple words from the deepest single layer rather than one word (ablation) | ELITE global mapping · DreamBench | DINO-I: 0.546 -> 0.431 (full embeddings) | Hurt |
| add local mapping network (ablation) | ELITE · DreamBench | CLIP-I / DINO-I: CLIP-I 0.699 -> 0.762; DINO-I 0.486 -> 0.652 | Improved |
| add local mapping network (ablation) | ELITE · DreamBench | CLIP-T: 0.257 -> 0.255 | No clear effect |
| increase fusion weight lambda of local attention from 0 to 1.2 (ablation) | ELITE local mapping · DreamBench | CLIP-I, DINO-I, CLIP-T | Mixed |
| place deepest layer word first versus middle or last among layer words (ablation) | ELITE global mapping · DreamBench | CLIP-T: 0.249 / 0.247 -> 0.257 | Improved |
| add local attention map reweighting by primary word attention (ablation) | ELITE local mapping · qualitative | editability | Improved |
| add global mapping network to local-only mapping (ablation) | ELITE · qualitative | concept fidelity | Improved |
| replace per-concept optimization with feed-forward encoder | customized text-to-image generation · DreamBench | encoding time: 6 min -> 0.05s (vs Custom Diffusion) | Improved |
| use ELITE instead of Custom Diffusion / DreamBooth | customized generation · DreamBench | CLIP-I / DINO-I: CLIP-I 0.801 -> 0.762; DINO-I 0.695 -> 0.652 | Hurt |
Later papers that compared with this method
15 later accepted papers report ELITE as a baseline in their results tables; 15 of them report a better number on most of the results they share with it.
- ConceptPrism: Concept Disentanglement in Personalized Diffusion Models via Residual Token Optimization (CVPR 2026 · better on 3 of 3 shared results)
- TOPA: Mitigating Concept Dominance in Diffusion Personalization via Target-Oriented Perturbation Augmentation (ECCV 2026 · better on 2 of 2 shared results)
- UniversalBooth: Model-Agnostic Personalized Text-to-Image Generation (ICCV 2025 · better on 8 of 9 shared results)
- Harmonizing Visual and Textual Embeddings for Zero-Shot Text-to-Image Customization (AAAI 2025 · better on 8 of 8 shared results)
- ObjectMate: A Recurrence Prior for Object Insertion and Subject-Driven Generation (ICCV 2025 · better on 2 of 2 shared results)
- Large-Scale Text-to-Image Model with Inpainting is a Zero-Shot Subject-Driven Image Generator (CVPR 2025 · better on 1 of 1 shared result)
- DisEnvisioner: Disentangled and Enriched Visual Prompt for Customized Image Generation (ICLR 2025 · better on 1 of 1 shared result)
- InstantBooth: Personalized Text-to-Image Generation without Test-Time Finetuning (CVPR 2024 · better on 5 of 7 shared results)
7 more not listed.
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
- Decoupled Textual Embeddings for Customized Image Generation (AAAI 2024 · similarity 0.89)
- An Image is Worth Multiple Words: Discovering Object Level Concepts using Multi-Concept Prompt Learning (ICML 2024 · similarity 0.88)
- AttnDreamBooth: Towards Text-Aligned Personalized Text-to-Image Generation (NeurIPS 2024 · similarity 0.88)
- Language-Informed Visual Concept Learning (ICLR 2024 · similarity 0.87)
- Salient Concept-Aware Generative Data Augmentation (NeurIPS 2025 · similarity 0.87)
- MC^2: Multi-concept Guidance for Customized Multi-concept Generation (CVPR 2025 · similarity 0.87)
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.
