News
2021-10-18 The Object Drawer, a high-performing end-to-end 3D object reconstruction system, will be available.
2021-04-21 The refined 3D-FRONT dataset has been released.
2021-03-19 The Trescope, a light-weight rendering tool, has been released.
2020-11-19 Release of report for 3D-FRONT dataset. If you use our dataset or code, please consider to cite our reports.
2020-06-14 Release of 3D-FRONT dataset announced at CVPR 2020 Workshop on Learning 3D Generative Models.
Introduction
We introduce 3D-FRONT (3D Furnished Rooms with layOuts and semaNTics), a new, large-scale, and comprehensive repository of synthetic indoor scenes highlighted by professionally designed layouts and a large number of rooms populated by high-quality textured 3D models with style compatibility. From layout semantics down to texture details of individual objects, our dataset is freely available to the academic community and beyond. Currently, 3D-FRONT contains 18,797 rooms diversely furnished by 3D objects, far surpassing all publicly available scene datasets. In addition, the 7,302 furniture objects all come with high-quality textures. While the floorplans and layout designs are directly sourced from professional creations, the interior designs in terms of furniture styles, color, and textures have been carefully curated based on a recommender system we develop to attain consistent styles as expert designs. Furthermore, we release Trescope, a light-weight rendering tool, to support benchmark rendering of 2D images and annotations from 3D-FRONT. We demonstrate two applications, interior scene synthesis and texture synthesis, that are especially tailored to the strengths of our new dataset.
Features
Alibaba will continously enlarge the benchmark by providing more annotations and adding new features (e.g., sharing the professional designs) to serve for 3D acadamic studies.
Exquisite interior design, complete 3D house layout
The most compelling part of 3D-FRONT is the indoor scene packages. 3D-FRONT shares all the essential data that constructs a scene, from layout semantics down to stylistic and texture details of individual objects. While the layout ideas are directly sourced from professional designs, the interior designs are transferred from expert creations followed by a post verification process. With such a feature, its scale, including 6,813 distinct houses and 18,797 diversely furnished rooms, is far surpassing all publicly available scene datasets. It enables data-driven designing studies, such as floorplans synthesis, interior scene synthesis, and scene suits compatibility prediction, that other scene datasets can not support well. It also benefits the studying of 3D scene understanding subjects, such as SLAM, 3D scene reconstruction, and 3D scene segmentation. Further, we assign selected practical camera viewpoints to furnished scenes and release the Trescope light-weight render tool. This allows the users of 3D-FRONT to easily render images and annotations to support their 2D vision studies. Last but not least, we promise that we will continuously improve 3D-FRONT by adding more features. Some certain plans include 1) releasing an industrial render engine, i.e., AceRay, to support benchmark photo-realistic rendering and 2) sharing much enriched texture and 3D geometry contents.
Download
The 3D-FRONT benchmark is provided by Alibaba Topping Homestyler, and organized by Alibaba Tao Technology Department under the 3D-FRONT Terms of Use. If you would like to download the 3D-FRONT data, please apply for the dateset at TC Lab: 3D-FRONT Dataset. If you have any questions, please send email to us at tianchi_bigdata@member.alibaba.com
.
Citation
We share the dataset in advance to facilitate related 3D gemotery and vision studies. If you use our dataset or code, please also cite our reports:
@inproceedings{fu20213d,
title={3d-front: 3d furnished rooms with layouts and semantics},
author={Fu, Huan and Cai, Bowen and Gao, Lin and Zhang, Ling-Xiao and Wang, Jiaming and Li, Cao and Zeng, Qixun and Sun, Chengyue and Jia, Rongfei and Zhao, Binqiang and others},
booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
pages={10933--10942},
year={2021}
}
@article{fu20213d,
title={3d-future: 3d furniture shape with texture},
author={Fu, Huan and Jia, Rongfei and Gao, Lin and Gong, Mingming and Zhao, Binqiang and Maybank, Steve and Tao, Dacheng},
journal={International Journal of Computer Vision},
pages={1--25},
year={2021},
publisher={Springer}
}