An Empty Room is All We Want: Automatic Defurnishing of Indoor Panoramas

arXiv:2405.03682v1 Announce Type: new
Abstract: We propose a pipeline that leverages Stable Diffusion to improve inpainting results in the context of defurnishing -- the removal of furniture items from indoor panorama images. Specifically, we illustrate how increased context, domain-specific model fine-tuning, and improved image blending can produce high-fidelity inpaints that are geometrically plausible without needing to rely on room layout estimation. We demonstrate qualitative and quantitative improvements over other furniture removal techniques.

abstract arxiv context cs.cv diffusion domain empty fidelity fine-tuning furniture image images inpainting model fine-tuning panorama pipeline results room stable diffusion type