Align3D: Image-Structure Co-Aligned 3D Scene Generation with Observability-Aware Fidelity-Preserving Guidance

A new arXiv paper (2611.08861v1) presents Align3D, an image-conditioned diffusion model that generates 3D indoor layouts faithful to a single input photograph while remaining physically plausible beyond the observed view. Align3D infers a confidence-calibrated scene graph from the photograph and trains a set-diffusion layout denoiser with image-structure co-alignment, using confidence-weighted modality fusion and observability-aware fidelity-preserving guidance that reweights conflicting scene-quality gradients with an object-level confidence mask. On 3D-FRONT, Align3D improves image-alignment and scene-quality metrics over recent baselines and maps previously unseen photographs to layouts in a single feed-forward pass once trained.

arXiv:2611.08861v1 Announce Type: new Abstract: We present Align3D, an image-conditioned diffusion model that generates 3D indoor layouts that are both faithful to a single input photograph and physically plausible beyond the observed view. From the photograph we infer a confidence-calibrated scene graph and train a set-diffusion layout denoiser with image-structure co-alignment where confidence-weighted modality fusion balances image and scene-graph alignment using predicted confidences, while observability-aware fidelity-preserving guidance projects conflicting scene-quality gradients and reweighs them with an object-level confidence mask so under-observed degrees of freedom absorb physical priors without undoing the image-graph match. On 3D-FRONT, Align3D improves image-alignment and scene-quality metrics over recent baselines; once trained, it maps previously unseen photographs to layouts in a single feed-forward pass.