Structured3D: A Large Photo-realistic Dataset for Structured 3D Modeling

Structured3D is a large-scale photo-realistic dataset containing 3.5K house designs (a) created by professional designers with a variety of ground truth 3D structure annotations (b) and generate photo-realistic 2D images (c).


  • 2020-07-03: The Structured3D dataset is accepted to ECCV 2020!
  • 2020-05-22: We are hosting the Holistic 3D Vision Challenges on the Holistic Scene Structures for 3D Vision Workshop at ECCV 2020.
  • 2019-10-16: The 3D bounding box of each instance is now available!
  • 2019-09-11: The perspective part of the Structured3D dataset is now available!
  • 2019-08-01: Structured3D dataset (panoramic images) and basic code for visualization are now available!


Recently, there has been growing interest in developing learning-based methods to detect and utilize salient semi-global or global structures, such as junctions, lines, planes, cuboids, smooth surfaces, and all types of symmetries, for 3D scene modeling and understanding. However, the ground truth annotations are often obtained via human labor, which is particularly challenging and inefficient for such tasks due to the large number of 3D structure instances (e.g., line segments) and other factors such as viewpoints and occlusions. In this paper, we present a new synthetic dataset, Structured3D, with the aim of providing large-scale photo-realistic images with rich 3D structure annotations for a wide spectrum of structured 3D modeling tasks. We take advantage of the availability of professional interior designs and automatically extract 3D structures from them. We generate high-quality images with an industry-leading rendering engine. We use our synthetic dataset in combination with real images to train deep networks for room layout estimation and demonstrate improved performance on benchmark datasets.


Structured3D: A Large Photo-realistic Dataset for Structured 3D Modeling
Jia Zheng*, Junfei Zhang*, Jing Li, Rui Tang, Shenghua Gao, Zihan Zhou
European Conference on Computer Vision (ECCV), 2020
[pdf] [supp] [code]
(* Equal contribution)


To download the dataset, please fill the agreement form that indicates you agree to the Structured3D Terms of Use. After we receive your agreement form, we will provide download access to the dataset. We also provide the basic code for viewing the structure annotations of our dataset.


The data is released under the Structured3D Terms of Use, and the code is released under the MIT license.


Jia Zheng
ShanghaiTech University
Junfei Zhang
Jing Li
ShanghaiTech University
Rui Tang
Shenghua Gao
ShanghaiTech University
Zihan Zhou
Penn State University

Related Links

  • Ctrl-Room provides the semantic labels for 3D bounding box annotations.
  • PNVS provides more panoramic views for novel view synthesis task.


We would like to thank for providing the database of house designs and the rendering engine. We especially thank Qing Ye and Qi Wu from for their help with the data rendering.