Research map / Image generation / Text-to-image diffusion
LACONIC: A 3D Layout Adapter for Controllable Image Creation
Léopold Maillard, Tom Durand, Adrien Ramanana Rahary, Maks Ovsjanikov
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 a pretrained-T2I adapter instead of training a diffusion model from scratch | LACONIC layout encoder · HyperSim (w/o text prompt) | FID: 15.83 -> 9.50 | Improved |
| use adapter instead of from-scratch model | LACONIC · HyperSim (w/o text prompt) | IS: 8.69 -> 9.74 | Improved |
| use adapter instead of from-scratch model | LACONIC · HyperSim (w/o text prompt) | KID x1000: 7.29 -> 3.44 | Improved |
| use explicit 3D layout adapter instead of SceneCraft depth/semantic ControlNets | LACONIC · HyperSim (w/o text prompt) | FID: 39.36 -> 9.50 | Improved |
| use explicit 3D layout adapter instead of SceneCraft | LACONIC · HyperSim (w/ text prompt) | FID: 27.69 -> 10.12 | Improved |
| use explicit 3D layout adapter instead of SceneCraft | LACONIC · HyperSim (w/ text prompt) | Inception Score: 14.55 -> 10.60 | Hurt |
| use explicit 3D layout adapter instead of SceneCraft | LACONIC · HyperSim (w/ text prompt) | SOC: 17.40 -> 18.39 | Improved |
| add world-to-camera coordinate reframing of the layout (ablation) | LACONIC · Custom bedroom dataset | FID: 34.03 -> 9.68 | Improved |
| add world-to-camera coordinate reframing of the layout (ablation) | LACONIC · Custom bedroom dataset | KID: 32.17 -> 10.24 | Improved |
| add world-to-camera coordinate reframing of the layout (ablation) | LACONIC · Custom bedroom dataset | SOC: 22.17 -> 22.35 | Improved |
| add transformer encoder in the layout encoder (ablation) | LACONIC · Custom bedroom dataset | FID: 10.68 -> 9.68 | Improved |
| add transformer encoder in the layout encoder (ablation) | LACONIC · Custom bedroom dataset | KID: 11.19 -> 10.24 | Improved |
Closest papers by meaning
- Build-A-Scene: Interactive 3D Layout Control for Diffusion-Based Image Generation (ICLR 2025 · similarity 0.90)
- SceneFactor: Factored Latent 3D Diffusion for Controllable 3D Scene Generation (CVPR 2025 · similarity 0.88)
- Generating compositional scenes via Text-to-image RGBA Instance Generation (NeurIPS 2024 · similarity 0.88)
- $\textit{Bifr\"ost}$: 3D-Aware Image Compositing with Language Instructions (NeurIPS 2024 · similarity 0.88)
- VD3D: Taming Large Video Diffusion Transformers for 3D Camera Control (ICLR 2025 · similarity 0.87)
- Thinking in Boxes: 3D Editing in Real Images Made Easy (NeurIPS 2026 · similarity 0.87)
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.
