论文标题
具有时间感知神经体素的快速动态辐射场
Fast Dynamic Radiance Fields with Time-Aware Neural Voxels
论文作者
论文摘要
神经辐射场（NERF）在建模3D场景和合成新视图图像方面取得了巨大成功。但是，大多数以前的NERF方法需要大量时间来优化一个场景。明确的数据结构，例如体素特征，显示出加速训练过程的巨大潜力。但是，体素特征面临两个大挑战，要应用于动态场景，即建模时间信息并捕获不同的点运动尺度。我们通过用时间感知的体素特征（称为Tineuvox）表示场景来提出一个辐射现场框架。引入了一个微小的坐标变形网络，以模拟粗线轨迹，并在辐射网络中进一步增强了时间信息。提出了一种多距离插值方法，并应用于体素特征，以模拟小动作和大型运动。我们的框架大大加速了动态光辉场的优化，同时保持高渲染质量。经验评估均在合成场景和真实场景上进行。我们的Tineuvox仅使用8分钟和8-MB的存储成本完成培训，同时表现出比以前的动态NERF方法相似甚至更好的渲染性能。
Neural radiance fields (NeRF) have shown great success in modeling 3D scenes and synthesizing novel-view images. However, most previous NeRF methods take much time to optimize one single scene. Explicit data structures, e.g. voxel features, show great potential to accelerate the training process. However, voxel features face two big challenges to be applied to dynamic scenes, i.e. modeling temporal information and capturing different scales of point motions. We propose a radiance field framework by representing scenes with time-aware voxel features, named as TiNeuVox. A tiny coordinate deformation network is introduced to model coarse motion trajectories and temporal information is further enhanced in the radiance network. A multi-distance interpolation method is proposed and applied on voxel features to model both small and large motions. Our framework significantly accelerates the optimization of dynamic radiance fields while maintaining high rendering quality. Empirical evaluation is performed on both synthetic and real scenes. Our TiNeuVox completes training with only 8 minutes and 8-MB storage cost while showing similar or even better rendering performance than previous dynamic NeRF methods.
