AI-Driven Design and Generative Architecture

BlenderProc

Maximilian Denninger, Martin Sundermeyer, Dominik Winkelbauer, Youssef Zidan, Dmitry Olefir, Mohamad Elbadrawy, Ahsan Lodhi, Harinandan Katam

Published 2019-10-25

Thumbnail for BlenderProc
📄 Paper🌐 Project💻 Code
BlenderProc is a modular procedural pipeline, which helps in generating real looking images for the training of convolutional neural networks. These can be used in a variety of use cases including segmentation, depth, normal and pose estimation and many others. A key feature of our extension of blender is the simple to use modular pipeline, which was designed to be easily extendable. By offering standard modules, which cover a variety of scenarios, we provide a starting point on which new modules can be created.

Abstract

BlenderProc is a modular procedural pipeline, which helps in generating real looking images for the training of convolutional neural networks. These can be used in a variety of use cases including segmentation, depth, normal and pose estimation and many others. A key feature of our extension of blender is the simple to use modular pipeline, which was designed to be easily extendable. By offering standard modules, which cover a variety of scenarios, we provide a starting point on which new modules can be created.

Topics

3D ReconstructionDatasetFloor Plan Reconstruction and ParametrizationGenerative Design and Design Space ExplorationTool/Librarygeometry generationprocedural modeling

Cite this paper

@misc{denninger2019blenderproc,
  title = {BlenderProc},
  author = {Maximilian Denninger and Martin Sundermeyer and Dominik Winkelbauer and Youssef Zidan and Dmitry Olefir and Mohamad Elbadrawy and Ahsan Lodhi and Harinandan Katam},
  year = {2019},
  eprint = {1911.01911},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1911.01911}
}