Plan2Scene: Converting Floorplans to 3D Scenes

TL;DR

Plan2Scene converts floorplans and photos into textured 3D models by lifting 2D images to 3D, synthesizing textures, and inferring unobserved surfaces using neural networks, enabling realistic indoor scene reconstruction.

Contribution

The paper introduces a novel system that transforms floorplans and photos into textured 3D models, including new datasets and a graph neural network for texture inference.

Findings

Outperforms baseline methods on texture quality metrics.

Produces realistic 3D interior models.

Handles sparse, unaligned photo inputs effectively.

Abstract

We address the task of converting a floorplan and a set of associated photos of a residence into a textured 3D mesh model, a task which we call Plan2Scene. Our system 1) lifts a floorplan image to a 3D mesh model; 2) synthesizes surface textures based on the input photos; and 3) infers textures for unobserved surfaces using a graph neural network architecture. To train and evaluate our system we create indoor surface texture datasets, and augment a dataset of floorplans and photos from prior work with rectified surface crops and additional annotations. Our approach handles the challenge of producing tileable textures for dominant surfaces such as floors, walls, and ceilings from a sparse set of unaligned photos that only partially cover the residence. Qualitative and quantitative evaluations show that our system produces realistic 3D interior models, outperforming baseline approaches on a suite of texture quality metrics and as measured by a holistic user study.