We propose a novel-view augmentation (NOVA) strategy to train NeRFs for photo-realistic 3D composition of dynamic objects in a static scene. Compared to prior work, our framework significantly reduces blending artifacts when inserting multiple dynamic objects into a 3D scene at novel views and times; achieves comparable PSNR without the need for additional ground truth modalities like optical flow; and overall provides ease, flexibility, and scalability in neural composition. Our codebase is on GitHub.
@article{arxiv.2308.12560,
title = {NOVA: NOvel View Augmentation for Neural Composition of Dynamic Objects},
author = {Dakshit Agrawal and Jiajie Xu and Siva Karthik Mustikovela and Ioannis Gkioulekas and Ashish Shrivastava and Yuning Chai},
journal= {arXiv preprint arXiv:2308.12560},
year = {2023}
}
Comments
Accepted for publication in ICCV Computer Vision for Metaverse Workshop 2023 (code is available at https://github.com/dakshitagrawal/NoVA)