Buildee:

A 3D Simulation Framework for Scene Exploration and Reconstruction with Understanding

CVPR 2025 Workshop on Synthetic Data for Computer Vision

Clémentin BoittiauxVincent Lepetit

LIGM, École Nationale des Ponts et Chaussées, IP Paris, CNRS, France

A Blender-based simulation framework for Python


Buildee is a 3D simulation framework designed to benchmark scene exploration, 3D reconstruction, and semantic segmentation tasks in both static and dynamic environments. Built as a Python module on top of Blender, Buildee leverages its advanced rendering capabilities to generate realistic RGB, depth, and semantic data while enabling 2D/3D point tracking and occlusion checking. Additionally, we provide a procedural generator for construction site environments and baseline methods for key computer vision tasks. Through Buildee, we establish a standardized platform for evaluating scene understanding algorithms in realistic settings.

Get started quickly


With Buildee, you can start your simulation in just a few lines of code. The following example shows how to:

  1. Load a Blender scene.
  2. Render RGB image, depth map, and segmentation map.
  3. Compute the current point cloud of the scene (dynamic points positions may change over time).
  4. Step forward the simulation to the next frame.
from buildee import Simulator

sim = Simulator(
    blend_file='file.blend',
    points_density=10.0,
    verbose=True
)

rgb, depth, labels = sim.render()

pcl, pcl_labels, pcl_visibility_mask = sim.compute_point_cloud(imshow=True)

sim.step_frame()

Acknowledgements


This project was funded by the European Union (ERC Advanced Grant explorer, Funding ID 101097259).

We would like to thank the creators of the following 3D models used in our procedural generator and animations:

Citation


If you use Buildee, please consider citing our paper.

@inproceedings{boittiaux2025buildee,
    title={Buildee: A 3D Simulation Framework for Scene Exploration and Reconstruction with Understanding},
    author={Cl\'ementin Boittiaux and Vincent Lepetit},
    booktitle={CVPR Workshop on Synthetic Data for Computer Vision},
    year={2025}
}

This page uses ACE's template.