ReplaceAnything3D: Text-Guided Object Replacement in 3D Scenes with Compositional Scene Representations
Abstract
We introduce ReplaceAnything3D model RAM3D, a novel method for 3D object replacement in 3D scenes based on users' text description. Given multi-view images of a scene, a text prompt describing the object to replace, and another describing the new object, our Erase-and-Replace approach can effectively swap objects in 3D scenes with newly generated content while maintaining 3D consistency across multiple viewpoints. We demonstrate the versatility of RAM3D by applying it to various realistic 3D scene types, showcasing results of modified objects that blend in seamlessly with the scene without impacting its overall integrity.
Cite
Text
Bartrum et al. "ReplaceAnything3D: Text-Guided Object Replacement in 3D Scenes with Compositional Scene Representations." Neural Information Processing Systems, 2024. doi:10.52202/079017-1539Markdown
[Bartrum et al. "ReplaceAnything3D: Text-Guided Object Replacement in 3D Scenes with Compositional Scene Representations." Neural Information Processing Systems, 2024.](https://mlanthology.org/neurips/2024/bartrum2024neurips-replaceanything3d/) doi:10.52202/079017-1539BibTeX
@inproceedings{bartrum2024neurips-replaceanything3d,
title = {{ReplaceAnything3D: Text-Guided Object Replacement in 3D Scenes with Compositional Scene Representations}},
author = {Bartrum, Edward and Nguyen-Phuoc, Thu and Xie, Chris and Li, Zhengqin and Khan, Numair and Avetisyan, Armen and Lanman, Douglas and Xiao, Lei},
booktitle = {Neural Information Processing Systems},
year = {2024},
doi = {10.52202/079017-1539},
url = {https://mlanthology.org/neurips/2024/bartrum2024neurips-replaceanything3d/}
}