Papers › LaCon: Late-Constraint Diffusion for Steerable Guided Image Synthesis
LaCon: Late-Constraint Diffusion for Steerable Guided Image Synthesis
Chang Liu, Rui Li, Kaidong Zhang, Xin Luo, Dong Liu
Diffusion models have demonstrated impressive abilities in generating photo-realistic and creative images. To offer more controllability for the generation process, existing studies, termed as early-constraint methods in this paper, leverage extra conditions and incorporate them into pre-trained diffusion models. Particularly, some of them adopt condition-specific modules to handle conditions separately, where they struggle to generalize across other conditions. Although follow-up studies present unified solutions to solve the generalization problem, they also require extra resources to implement, e.g., additional inputs or parameter optimization, where more flexible and efficient solutions are expected to perform steerable guided image synthesis. In this paper, we present an alternative paradigm, namely Late-Constraint Diffusion (LaCon), to simultaneously integrate various conditions into pre-trained diffusion models. Specifically, LaCon establishes an alignment between the external condition and the internal features of diffusion models, and utilizes the alignment to incorporate the target condition, guiding the sampling process to produce tailored results. Experimental results on COCO dataset illustrate the effectiveness and superior generalization capability of LaCon under various conditions and settings. Ablation studies investigate the functionalities of different components in LaCon, and illustrate its great potential to serve as an efficient solution to offer flexible controllability for diffusion models.
In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Conditional Text-to-Image Synthesis | COCO 2017 val | LCDG | FID | 20.27 | #1 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | LCDG (Color, evaluated under image palette) | CLIP Score | 0.2580 | #2 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | LCDG (Color, evaluated under image palette) | FID | 20.61 | #2 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | LCDG (Mask) | CLIP Score | 0.2617 | #3 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | LCDG (Mask) | FID | 20.94 | #3 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | LCDG (Edge) | FID | 21.02 | #4 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | T2I-Adapter (Sketch) | CLIP Score | 0.2580 | #5 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | T2I-Adapter (Sketch) | FID | 21.72 | #5 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | T2I-Adapter (Color, evaluated under image palette) | CLIP Score | 0.2613 | #6 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | T2I-Adapter (Color, evaluated under image palette) | FID | 26.54 | #6 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | SD (text) | CLIP Score | 0.2673 | #7 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | SD (text) | FID | 27.99 | #7 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | ControlNet (HED Edge) | CLIP Score | 0.2525 | #8 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | ControlNet (HED Edge) | FID | 28.09 | #8 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | T2I-Adapter (Color, evaluated under color stroke) | FID | 30.84 | #9 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | SD using SDEdit (evaluated under color stroke) | CLIP Score | 0.2257 | #10 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | SD using SDEdit (evaluated under color stroke) | FID | 32.93 | #10 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | SD using SDEdit | FID | 71.16 | #11 of 12 | Archive leaderboard | report |
| Conditional Text-to-Image Synthesis | COCO 2017 val | SD using SDEdit (evaluated under image palette) | CLIP Score | 0.2138 | #12 of 12 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Methods
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections