Linear Fusion MultiDiffusion for Fast Training-Free Spherical Panorama Generation

arXiv:2609.01997 · cs.CV, cs.LG · Submitted 2026-09-02 · Read on arXiv

cs.CV, cs.LG

Submitted: 2026-09-02

Updated: 2026-09-02

Comments: Accepted to ECCV 2026

Code: https://github.com/black-forest-labs/flux

Project page: https://ahykw.github.io/lfmd

License: http://creativecommons.org/licenses/by/4.0/

The gist: We propose LF-MultiDiffusion, a training-free panorama generation method that extends MultiDiffusion to support linear projections between target and reference image spaces.

Terminology

Abstract

We propose LF-MultiDiffusion, a training-free panorama generation method that extends MultiDiffusion to support linear projections between target and reference image spaces. Our key idea is to reformulate latent aggregation as a regularized least-squares problem and solve it efficiently with a Krylov-based iterative solver inside the denoising loop. This formulation enables denser and more natural mappings than prior training-free methods, yielding more stable generation with far fewer perspective views. As a result, LF-MultiDiffusion reduces the number of image generator evaluations during denoising and significantly improves inference efficiency. Experiments show that LF-MultiDiffusion achieves better visual quality, text alignment, and panoramic consistency than the strongest training-free baseline, while providing a 15.36 times speedup. Our project page is available at: https://ahykw.github.io/lfmd.

Related papers