3D Scene Painting via Semantic Image Synthesis
Abstract
We propose a novel approach to 3D scene painting using a configurable 3D scene layout. Our approach takes a 3D scene with semantic class labels as input and trains a 3D scene painting network that synthesizes color values for the input 3D scene. We exploit an off-the-shelf 2D semantic image synthesis method to teach the 3D painting network without explicit color supervision. Experiments show that our approach produces images with geometrically correct structures and supports scene manipulation, such as the change of viewpoint, object poses, and painting style. Our approach provides rich controllability to synthesized images in the aspect of 3D geometry.
Cite
Text
Jeong et al. "3D Scene Painting via Semantic Image Synthesis." Conference on Computer Vision and Pattern Recognition, 2022. doi:10.1109/CVPR52688.2022.00230Markdown
[Jeong et al. "3D Scene Painting via Semantic Image Synthesis." Conference on Computer Vision and Pattern Recognition, 2022.](https://mlanthology.org/cvpr/2022/jeong2022cvpr-3d/) doi:10.1109/CVPR52688.2022.00230BibTeX
@inproceedings{jeong2022cvpr-3d,
title = {{3D Scene Painting via Semantic Image Synthesis}},
author = {Jeong, Jaebong and Jo, Janghun and Cho, Sunghyun and Park, Jaesik},
booktitle = {Conference on Computer Vision and Pattern Recognition},
year = {2022},
pages = {2262-2272},
doi = {10.1109/CVPR52688.2022.00230},
url = {https://mlanthology.org/cvpr/2022/jeong2022cvpr-3d/}
}