Research map / Image generation / Text-to-image diffusion
Semantic Alignment for Pose-Invariant Identity Preserving Diffusion
Jiwon Kim, SeonHwa Kim, Soobin Park, Eunju Cha, Kyong Hwan Jin
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 |
|---|---|---|---|
| use SeAl identity infusion vs best training-free baseline | SeAl · Human Pose | DINO-I: 0.5588 (Ctrl-X) -> 0.7465 | Improved |
| use SeAl vs DRF | SeAl · Animal Pose | DINO-I: 0.6318 -> 0.7104 | Improved |
| use SeAl vs ControlNet+IP-Adapter | SeAl · Human Canny | DINO-I: 0.7140 -> 0.7887 | Improved |
| use SeAl vs Uni-ControlNet on text alignment | SeAl · Animal Depth | CLIP score: 0.3489 -> 0.3260 | Hurt |
| use SeAl vs T2I-Adapter+IP-Adapter on structure | SeAl · Human Mesh | DreamSim-str: 0.7004 -> 0.6300 | Hurt |
| add AnchorAlign pre-alignment (ablation) | SeAl · qualitative misaligned inputs | structure coherence | Improved |
| add AnchorAlign pre-alignment (ablation) | SeAl · inference cost | latency (s): 24.56 -> 34.83 | Hurt |
| increase top-k from 1 to larger values (ablation) | SeAl · qualitative | image sharpness | Hurt |
| increase semantic threshold tau (ablation) | SeAl · appearance transfer | DINO-I | Hurt |
| increase blending strength delta (ablation) | SeAl · appearance transfer | DINO-I | Improved |
| use SeAl vs baselines on reward models | SeAl · HPSv2 / ImageReward | reward score | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- Ctrl-X — Ctrl-X: Controlling Structure and Appearance for Text-To-Image Generation Without Guidance (NeurIPS 2024) · better on 6 of 7 shared results
- Uni-ControlNet — Uni-ControlNet: All-in-One Control to Text-to-Image Diffusion Models (NeurIPS 2023) · better on 0 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
- WithAnyone: Toward Controllable and ID Consistent Image Generation (ICLR 2026 · similarity 0.90)
- When StyleGAN Meets Stable Diffusion: a W+ Adapter for Personalized Image Generation (CVPR 2024 · similarity 0.90)
- TopoRefine: Topology-Aware Correspondence and Residual Refinement for Training-Free Subject-Consistent Generation (NeurIPS 2026 · similarity 0.90)
- FaithfulFaces: Pose-Faithful Facial Identity Preservation for Text-to-Video Generation (NeurIPS 2026 · similarity 0.88)
- DreamSalon: A Staged Diffusion Framework for Preserving Identity-Context in Editable Face Generation (CVPR 2024 · similarity 0.87)
- REPA Works Until It Doesn’t: Early-Stopped, Holistic Alignment Supercharges Diffusion Training (NeurIPS 2025 · similarity 0.86)
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.
