arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.01226cs.CVcs.GRcs.LG
收藏

Factor Fields:神经场及其扩展的统一框架

Factor Fields: A Unified Framework for Neural Fields and Beyond

ETH Zürich(苏黎世联邦理工学院);University of Tübingen(图宾根大学);Adobe Research(奥多比研究院);University of California, San Diego(加利福尼亚大学圣迭戈分校);Tübingen AI Center(图宾根人工智能中心)

Anpei Chen, Zexiang Xu, Xinyue Wei, Siyu Tang, Hao Su, Andreas Geiger

AI总结:

本文提出Factor Fields统一框架,将信号分解为因子乘积以整合NeRF等现有神经场表示,并基于此提出新表示DiF,在近似质量、紧凑性与训练效率上取得显著提升,且支持少样本泛化。

摘要:

我们提出了Factor Fields,一种用于建模和表示信号的新颖框架。Factor Fields将信号分解为多个因子的乘积,每个因子由作用于变换后输入坐标的经典或神经场表示来表示。这种分解形成了一个统一框架,能够容纳近期多种信号表示方法,包括NeRF、Plenoxels、EG3D、Instant-NGP和TensoRF。此外,该框架允许构建强大的新信号表示,例如本文的第二个贡献“Dictionary Field”(DiF)。实验表明,与先前的快速重建方法相比,DiF在近似质量、紧凑性和训练时间方面均有提升。在实验中,该表示在2D图像回归任务中实现了更好的图像近似质量,在重建3D符号距离场时具有更高的几何质量,并在辐射场重建任务中展现出更高的紧凑性。此外,DiF通过在训练期间跨信号共享基来实现对未见图像/3D场景的泛化,这极大地有益于稀疏观测下的图像回归和少样本辐射场重建等用例。

英文摘要:

We present Factor Fields, a novel framework for modeling and representing signals. Factor Fields decomposes a signal into a product of factors, each represented by a classical or neural field representation which operates on transformed input coordinates. This decomposition results in a unified framework that accommodates several recent signal representations including NeRF, Plenoxels, EG3D, Instant-NGP, and TensoRF. Additionally, our framework allows for the creation of powerful new signal representations, such as the "Dictionary Field" (DiF) which is a second contribution of this paper. Our experiments show that DiF leads to improvements in approximation quality, compactness, and training time when compared to previous fast reconstruction methods. Experimentally, our representation achieves better image approximation quality on 2D image regression tasks, higher geometric quality when reconstructing 3D signed distance fields, and higher compactness for radiance field reconstruction tasks. Furthermore, DiF enables generalization to unseen images/3D scenes by sharing bases across signals during training which greatly benefits use cases such as image regression from sparse observations and few-shot radiance field reconstruction.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑