Research map / Image generation / Text-to-image diffusion
Enhancing MMDiT-Based Text-to-Image Models for Similar Subject Generation
Tianyi Wei, Dongdong Chen, Yifan Zhou, Xingang Pan
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 block alignment loss (ablation) | MMDiT latent optimization (SD3) · three similar subjects | SR (Grounding DINO): +8.3% | Improved |
| add overlap loss (ablation) | MMDiT latent optimization (SD3) · three similar subjects | SR (Grounding DINO): +8.6% | Improved |
| add text encoder alignment loss to form L_amb (ablation) | MMDiT latent optimization (SD3) · three similar subjects | SR (Grounding DINO): 57.3% | Improved |
| add overlap online detection and back-to-start sampling (ablation) | L_amb (SD3) · three similar subjects | SR (Grounding DINO): 57.3% -> 66.7% | Improved |
| add reject sampling (ablation) | SD3 with L_amb + L_res · three similar subjects | SR (Grounding DINO): 66.7% -> 67.8% | Improved |
| add reject sampling (ablation) | SD3 with L_amb + L_res · four similar subjects | SR (Grounding DINO): 52.7% -> 60.8% | Improved |
| add reject sampling (ablation) | SD3 with L_amb + L_res · two similar subjects | SR (Grounding DINO): 68.8% -> 68.8% | No clear effect |
| reverse block alignment constraint (later blocks constrained to earlier) (ablation) | block alignment loss · similar subjects | generation quality | Hurt |
| use our method instead of CONFORM | SD3 Medium · four similar subjects | SR (Grounding DINO): 43.7% -> 68.7% (GPT-4o) | Improved |
| apply method on SD3.5 instead of SD3 | SD3.5 Medium · four similar subjects | SR (Grounding DINO): 26.2% -> 65.9% | Improved |
| add back-to-start sampling | L_amb · two similar subjects | sampling time: about 68.7 seconds; triggered 0.4%, 29.3%, 42.1% | Mixed |
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.
- CONFORM — CONFORM: Contrast is All You Need for High-Fidelity Text-to-Image Diffusion Models (CVPR 2024) · better on 9 of 9 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
- Rare Text Semantics Were Always There in Your Diffusion Transformer (NeurIPS 2025 · similarity 0.89)
- Exploring the Deep Fusion of Large Language Models and Diffusion Transformers for Text-to-Image Synthesis (CVPR 2025 · similarity 0.89)
- Information Theoretic Text-to-Image Alignment (ICLR 2025 · similarity 0.88)
- Improving Long-Text Alignment for Text-to-Image Diffusion Models (ICLR 2025 · similarity 0.88)
- Dual Diffusion for Unified Image Generation and Understanding (CVPR 2025 · similarity 0.88)
- AP-Adapter: Improving Generalization of Automatic Prompts on Unseen Text-to-Image Diffusion Models (NeurIPS 2024 · 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.
