Research map / Image generation / Text-to-image diffusion
ControlNet++: Improving Conditional Controls with Efficient Consistency Feedback
Ming Li, Taojiannan Yang, Huafeng Kuang, Jie Wu, Zhaoning Wang, Xuefeng Xiao, Chen Chen
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 pixel-level consistency reward fine-tuning to ControlNet | ControlNet (SD1.5) · ADE20K (seg mask condition) | mIoU: 32.55 -> 43.64 | Improved |
| add consistency reward fine-tuning | ControlNet (SD1.5) · MultiGen-20M depth map | RMSE: 35.90 -> 28.32 | Improved |
| add consistency reward fine-tuning | ControlNet (SD1.5) · MultiGen-20M canny edge | FID: 14.73 -> 18.23 | Hurt |
| add consistency reward fine-tuning | ControlNet (SD1.5) · ADE20K, COCO, hed, lineart, depth | FID: 33.28 -> 29.49 (ADE20K) | Improved |
| add consistency reward fine-tuning | ControlNet (SD1.5) · CLIP score (several conditions) | CLIP score: 32.15 -> 31.87 (canny) | No clear effect |
| use a stronger text-to-image backbone (SDXL instead of SD1.5) | ControlNet / T2I-Adapter · MultiGen-20M depth map | RMSE: 35.90 (SD1.5) vs 40.00 (SDXL) | No clear effect |
| add real data augmented with ControlNet++ generated images for segmentation training | DeepLabv3-MBv2 on ADE20K · ADE20K | mIoU: +1.76 mIoU | Improved |
| augment real data with ControlNet-generated images | DeepLabv3-MBv2 on ADE20K · ADE20K | mIoU: 34.11 vs 34.08 | No clear effect |
| train with consistency loss together with the diffusion loss rather than consistency loss only (ablation) | ControlNet++ · ADE20K | image quality / controllability | Improved |
| fine-tune only a subset of timesteps with reward, evaluated by crossing models across timesteps (ablation) | ControlNet++ · ADE20K | mIoU: 32.55 -> 38.03 | Improved |
| use a stronger reward model (UperNet-R50 instead of DeepLabv3-MBv2) (ablation) | ControlNet++ · ADE20K | mIoU: 31.96 -> 43.64 | Improved |
| use the weakest reward model (DeepLabv3-MBv2) (ablation) | ControlNet++ · ADE20K | mIoU: 32.55 -> 31.96 | Hurt |
Later papers that compared with this method
77 later accepted papers report ControlNet as a baseline in their results tables; 65 of them report a better number on most of the results they share with it.
- Gated Condition Injection without Multimodal Attention: Towards Controllable Linear-Attention Transformers (CVPR 2026 · better on 8 of 10 shared results)
- SketchEvo: Leveraging Drawing Dynamics for Enhanced Image Synthesis (ICLR 2026 · better on 6 of 7 shared results)
- Text-guided Controllable Diffusion for Realistic Camouflage Images Generation (AAAI 2026 · better on 5 of 5 shared results)
- MMFace-DiT: A Dual-Stream Diffusion Transformer for High-Fidelity Multimodal Face Generation (CVPR 2026 · better on 4 of 4 shared results)
- Reflection Separation from a Single Image via Joint Latent Diffusion (CVPR 2026 · better on 3 of 3 shared results)
- Semantic Context Matters: Improving Conditioning for Autoregressive Models (CVPR 2026 · better on 3 of 3 shared results)
- BideDPO: Conditional Image Generation with Simultaneous Text and Condition Alignment (ICLR 2026 · better on 2 of 2 shared results)
- Controllable diffusion-based generation for multi-channel biological data (ICLR 2026 · better on 2 of 2 shared results)
69 more not listed.
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.
- UniControl — UniControl: A Unified Diffusion Model for Controllable Visual Generation In the Wild (NeurIPS 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
- Simplifying Control Mechanism in Text-to-Image Diffusion Models (AAAI 2025 · similarity 0.93)
- CtrLoRA: An Extensible and Efficient Framework for Controllable Image Generation (ICLR 2025 · similarity 0.92)
- CPO: Condition Preference Optimization for Controllable Image Generation (NeurIPS 2025 · similarity 0.91)
- Local Conditional Controlling for Text-to-Image Diffusion Models (AAAI 2025 · similarity 0.91)
- FlexControl: Computation-Aware Conditional Control with Differentiable Router for Text-to-Image Generation (ICML 2025 · similarity 0.91)
- ControlAR: Controllable Image Generation with Autoregressive Models (ICLR 2025 · similarity 0.90)
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.
