Neural radiance field (NeRF) has shown remarkable performance in generating photo-realistic novel views. Since the emergence of NeRF, many studies have been conducted, among which managing features with explicit structures such as grids has achieved exceptionally fast training by reducing the complexity of multilayer perceptron (MLP) networks. However, storing features in dense grids requires significantly large memory space, which leads to memory bottleneck in computer systems and thus large training time. To address this issue, in this work, we propose MixNeRF, a memory-efficient NeRF framework that employs a mixed-up hash table to improve memory efficiency and reduce training time while maintaining reconstruction quality. We first design a mixed-up hash table to adaptively mix part of multi-level feature grids into one and map it to a single hash table. Following that, in order to obtain the correct index of a grid point, we further design an index transformation method that transforms indices of an arbitrary level grid to those of a canonical grid. Extensive experiments benchmarking with state-of-the-art Instant-NGP, TensoRF, and DVGO, indicate our MixNeRF could achieve the fastest training time on the same GPU hardware with similar or even higher reconstruction quality. Source code is available at https://github.com/nfyfamr/MixNeRF.

翻译：神经辐射场（NeRF）在新视图合成中展现出生成逼真图像的卓越性能。自NeRF问世以来，大量研究随之展开，其中通过网格等显式结构管理特征的方法因降低了多层感知机（MLP）网络的复杂度而实现了极快的训练速度。然而，在密集网格中存储特征需要占用巨大的内存空间，这导致计算机系统出现内存瓶颈，进而延长训练时间。为解决该问题，本文提出MixNeRF——一种内存高效的NeRF框架，它采用混合哈希表来提升内存效率、减少训练时间，同时保持重建质量。我们首先设计了一个混合哈希表，能够自适应地将部分多层特征网格混合为一个，并将其映射至单一哈希表；随后，为获取网格点的正确索引，我们进一步设计了一种索引变换方法，可将任意层级网格的索引转换为规范网格的索引。与当前最先进的Instant-NGP、TensoRF和DVGO进行的大量对比实验表明，在相同GPU硬件条件下，MixNeRF能以相似甚至更高的重建质量实现最快的训练时间。源代码已开源至https://github.com/nfyfamr/MixNeRF。