Research map / Image generation / Text-to-image diffusion
ST-LDM: A Universal Framework for Text-Grounded Object Generation in Real Images
Xiangtian Xue, Jiasong Wu, Youyong Kong, Lotfi Senhadji, Huazhong Shu
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 |
|---|---|---|---|
| apply deformable feature alignment in stages 2-4 versus no DFA (ablation) | ST-LDM multimodal transformer · TOG benchmark | FID: 27.03 -> 6.16 | Improved |
| also apply DFA in the initial stage 1 (ablation) | ST-LDM · TOG benchmark | FID: 6.16 -> 7.50 | Hurt |
| add backward guidance with guidance map (ablation) | ST-LDM · TOG benchmark | UserScore: 58.73 -> 89.11 | Improved |
| add threshold activation to guidance map (ablation) | ST-LDM · TOG benchmark | UserScore: 80.86 -> 89.11 | Improved |
| add dilation after activation (ablation) | ST-LDM · TOG benchmark | UserScore: 83.98 -> 89.11 | Improved |
| add learned 2D offsets in DFA (ablation) | DFA · TOG benchmark | IoU: 49.88 -> 65.12 | Improved |
| add modulation scalar in DFA (ablation) | DFA · TOG benchmark | IoU: 52.49 -> 65.12 | Improved |
| combine transformer layer with multi-head cross attention in DFA (ablation) | DFA · TOG benchmark | IoU: 28.95 (T-only) / 32.86 (C-only) -> 65.12 | Improved |
| add ST-LDM guidance to Stable Diffusion | SD V-1.5 · inference time | seconds per image: ~4.1 -> ~6.5 s/img | Hurt |
| use ST-LDM with text only versus GLIGEN with bounding boxes | text-grounded object generation · TOG benchmark | FID: 6.24 -> 6.16 | No clear effect |
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.
- InstructPix2Pix — InstructPix2Pix: Learning To Follow Image Editing Instructions (CVPR 2023) · 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
- R&B: Region and Boundary Aware Zero-shot Grounded Text-to-image Generation (ICLR 2024 · similarity 0.90)
- Precise Parameter Localization for Textual Generation in Diffusion Models (ICLR 2025 · similarity 0.89)
- Progressive Text-to-Image Diffusion with Soft Latent Direction (AAAI 2024 · similarity 0.89)
- Localizing and Editing Knowledge In Text-to-Image Generative Models (ICLR 2024 · similarity 0.89)
- LLM Blueprint: Enabling Text-to-Image Generation with Complex and Detailed Prompts (ICLR 2024 · similarity 0.88)
- Brush Your Text: Synthesize Any Scene Text on Images via Diffusion Model (AAAI 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.
