【摘要】 神经辐射场是一种采用隐式场景表示的新视角合成算法，近年来在新视角合成及三维重建领域掀起了研究热潮。神经辐射场模型在城市重建、虚拟现实等方面都有重要应用。然而，用于训练神经辐射场的多视角图像在很多场景下往往难以获取，因此有必要回顾一下近两年关于稀疏输入视角神经辐射场的相关成果。论文介绍了基于神经辐射场的体渲染相关理论和近年来的新进展，并分类总结了解决稀疏输入问题的一些方法，希望能给后续的研究者提供一些参考。更多还原

【Abstract】 Neural radiance field(NeRF)is an implicit scene representation model,which has set off a research boom in the field of novel view synthesis and 3D reconstruction in recent years. The NeRF model has applications in urban reconstruction,virtual reality,etc. However,multi-view images for training are often difficult to obtain in many scenarios,so it is necessary to review the papers on sparse input views NeRF in the past two years. This paper introduces the related theory of NeRF-based neural volume rendering and some progress in recent years,and summarizes some methods to solve the sparse input views problem,hoping to give some inspiration to subsequent researchers.更多还原