VALSE 论文速览 第181期：ForwardDNeRF

论文题目：

Forward Flow for Novel View Synthesis of Dynamic Scenes

作者列表：

郭相 (西北工业大学), 孙家岱 (西北工业大学), 戴玉超 (西北工业大学), 陈冠英 (香港中文大学(深圳)), 叶晓青 (百度), 谭啸 (百度), 丁二锐 (百度), 张雨蒙 (百度), 王井东 (百度)

B站观看网址：

https://www.bilibili.com/video/BV1YZ421H7RC/

论文摘要：

This paper proposes a neural radiance field (NeRF) approach for novel view synthesis of dynamic scenes using forward warping. Existing methods often adopt a static NeRF to represent the canonical space, and render dynamic images at other time steps by mapping the sampled 3D points back to the canonical space with the learned \emph{backward flow｝ field. However, this backward flow field is non-smooth and discontinuous, which is difficult to be fitted by commonly used smooth motion models. To address this problem, we propose to estimate the \emph{forward flow｝ field and directly warp the canonical radiance field to other time steps. Such forward flow field is smooth and continuous within the object region, which benefits the motion model learning. To achieve this goal, we represent the canonical radiance field with voxel grids to enable efficient forward warping, and propose a differentiable warping process, including an average splatting operation and an inpaint network, to resolve the many-to-one and one-to-many mapping issues. Thorough experiments show that our method outperforms existing methods in both novel view rendering and motion modeling, demonstrating the effectiveness of our forward flow motion modeling.

论文链接：

[https://arxiv.org/abs/2309.17390]

视频讲者简介：

郭相是西北工业大学在读博士生，导师为戴玉超教授。2012年至2018年，在澳大利亚国立大学获得本科和硕士学位。目前在百度实习。主要研究方向为新视角生成和视觉定位，在ICCV，ACCV和ICPR发表多篇论文。新视角生成方面，主要的兴趣是关于动态场景的建模，训练加速，位姿优化等方向。视觉定位方面，主要的研究话题为基于隐式表达的视觉定位。

特别鸣谢本次论文速览主要组织者：

月度轮值AC：张瑞茂 (香港中文大学 (深圳))