OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Image & Video AI

Plan2Scene

Plan2Scene is an open-source Python research implementation that converts a residence floorplan and photos into a textured 3D mesh or scene. Its workflow includes texture synthesis for observed surfaces and texture propagation to unobserved surfaces.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Plan2Scene — OpenViST project profile in Image & Video AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Plan2Scene when a research or development task involves reconstructing a residential 3D scene from floorplan and photo inputs, including workflows based on Rent3D++ or custom floorplan data.

Who it is for

  • Researchers exploring indoor scene reconstruction from floorplans.
  • Developers working with floorplan and photo data who want to adapt or run the documented reconstruction workflow.

Documented uses

  • Create a textured 3D residence mesh from a floorplan and photos.
  • Propagate textures to surfaces not observed in the photos.
  • Prepare scene files with embedded textures for browser preview.
  • Train the texture synthesis and texture propagation components separately.
Before you choose

Requirements & limits

The repository is public and licensed under MIT. A Google Colab demo is linked; the documented local workflow requires a Conda environment and additional data and model assets.

  • Input data is central: the documented task uses both a floorplan and photos.
  • Custom scanned floorplans require conversion to vector format using Raster-to-Vector; other vector formats must be converted to its annotation format.
  • Local inference has separate environment, texture-synthesis, dataset, and model prerequisites.
  • Browser preview uses SmartScenesToolkit's scene-viewer, which must be cloned and built.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up Plan2Scene for local inference

Prepare the documented environment and dependencies for running the repository's inference workflow.

Before you start

  • A Conda environment initialized using the repository's environment setup instructions.
  • The Rent3D++ dataset copied to the project data directory for the documented dataset workflow.
  • A pretrained Plan2Scene model or a newly trained network.
  • The Embark Studios texture-synthesis command-line library, a seam mask, and the seam-correction configuration.

Steps

  1. Initialize the Conda environment using the repository's setup instructions.

  2. Install or build the texture-synthesis command-line library, download the seam mask, and configure their paths in seam_correct.json.

  3. Download and organize Rent3D++ in the project data directory, following the repository's data instructions.

  4. Set up a pretrained model or train a Plan2Scene network before proceeding to the repository's inference stages.

Expected outcome

The documented prerequisites for local Rent3D++ inference are in place; inference and output preview are separate subsequent workflows.

Guide limitations

  • A local run requires the dataset and model assets; this is a setup guide, not a claim that inference has been run.
Documentation 1 ↗Documentation 2 ↗
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
17 October 2026
Original website title
3dlg-hcvc/plan2scene · Official homepage title · retained in its source language
How we review sourcesSuggest a correction