Research map / Image generation / Fast diffusion sampling
From Open Loop to Closed Loop: A Test-Time Iterative Optimization Framework for Reference-Consistent Image Generation
Baixuan Zhao, Xinyu Zhang, 华渝 郑, Shuaicheng Liu, Xiongkuo Min, Guangtao Zhai, Xiaohong Liu
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 closed-loop feedback vs best-of-20 open loop | PuLID · CelebA300 | facial similarity: 0.560 -> 0.702 | Improved |
| add closed-loop feedback vs best-of-20 open loop | PMv2 · Web100 | facial similarity: 0.598 -> 0.684 | Improved |
| add closed-loop feedback vs best-of-20 open loop | InstantID · CelebA300 | facial similarity: 0.817 -> 0.878 | Improved |
| add closed-loop feedback vs best-of-20 open loop | IPA · Web100 | facial similarity: 0.684 -> 0.731 | Improved |
| add closed-loop feedback vs best-of-15 open loop | ControlNet · Pose-controlled generation | MPJPE: 36.521 -> 26.401 | Improved |
| add closed-loop feedback vs best-of-15 open loop | ControlNet · Depth-controlled generation | MAE: 15.797 -> 11.295 | Improved |
| add closed-loop feedback vs best-of-15 open loop | ControlNext · Pose-controlled generation | MPJPE: 75.827 -> 65.623 | Improved |
| add closed-loop feedback vs best-of-15 open loop | ControlNext · Depth-controlled generation | MAE: 29.598 -> 25.331 | Improved |
| add integral term to P controller (ablation) | Closed-loop PID · Identity (50 samples) | facial similarity: 0.694 -> 0.723 | Improved |
| add derivative term to PI controller (ablation) | Closed-loop PID · Identity (50 samples) | facial similarity: 0.723 -> 0.724 | No clear effect |
| add derivative term to P controller (ablation) | Closed-loop PID · Identity (50 samples) | iterations to 95% gain: 5.58 -> 5.52 | Improved |
| add closed-loop feedback with no text prompt (ablation) | Base generator · Prompt-level ablation, depth | MAE: 22.598 -> 14.553 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- PuLID — PuLID: Pure and Lightning ID Customization via Contrastive Alignment (NeurIPS 2024) · better on 2 of 2 shared results
- ControlNet — ControlNet++: Improving Conditional Controls with Efficient Consistency Feedback (ECCV 2024) · better on 2 of 2 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
- ControlAR: Controllable Image Generation with Autoregressive Models (ICLR 2025 · similarity 0.87)
- DC-ControlNet: Decoupling Inter- and Intra-Element Conditions in Image Generation with Diffusion Models (ICCV 2025 · similarity 0.87)
- FlexControl: Computation-Aware Conditional Control with Differentiable Router for Text-to-Image Generation (ICML 2025 · similarity 0.87)
- A Simple Approach to Unifying Diffusion-based Conditional Generation (ICLR 2025 · similarity 0.87)
- CtrLoRA: An Extensible and Efficient Framework for Controllable Image Generation (ICLR 2025 · similarity 0.86)
- CCS: Controllable and Constrained Sampling with Diffusion Models via Initial Noise Perturbation (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.
