Neural radiance fields (NeRFs) enable high-quality novel view synthesis, but their high computational complexity limits deployability. While existing neural-based solutions strive for efficiency, they use one-size-fits-all architectures regardless of scene complexity. The same architecture may be unnecessarily large for simple scenes but insufficient for complex ones. Thus, there is a need to dynamically optimize the neural network component of NeRFs to achieve a balance between computational complexity and specific targets for synthesis quality. We introduce NAS-NeRF, a generative neural architecture search strategy that generates compact, scene-specialized NeRF architectures by balancing architecture complexity and target synthesis quality metrics. Our method incorporates constraints on target metrics and budgets to guide the search towards architectures tailored for each scene. Experiments on the Blender synthetic dataset show the proposed NAS-NeRF can generate architectures up to 5.74$\times$ smaller, with 4.19$\times$ fewer FLOPs, and 1.93$\times$ faster on a GPU than baseline NeRFs, without suffering a drop in SSIM. Furthermore, we illustrate that NAS-NeRF can also achieve architectures up to 23$\times$ smaller, with 22$\times$ fewer FLOPs, and 4.7$\times$ faster than baseline NeRFs with only a 5.3% average SSIM drop. Our source code is also made publicly available at https://saeejithnair.github.io/NAS-NeRF.

翻译：神经辐射场（NeRFs）能够实现高质量的新视角合成，但其高计算复杂度限制了部署能力。尽管现有的基于神经网络的解决方案致力于提升效率，但它们采用了无视场景复杂度的通用架构。同一架构对于简单场景可能过于庞大，而对复杂场景则显不足。因此，需要动态优化NeRF的神经网络组件，以在计算复杂度与特定合成质量指标之间取得平衡。我们提出NAS-NeRF，一种生成式神经架构搜索策略，通过平衡架构复杂度与目标合成质量指标，生成紧凑且适配场景的NeRF架构。该方法将目标指标和计算预算作为约束条件，引导搜索面向每个场景定制的架构。在Blender合成数据集上的实验表明，与基线NeRF相比，所提出的NAS-NeRF可生成体积缩小5.74倍、浮点运算量减少4.19倍、GPU运行速度快1.93倍的架构，且SSIM指标未出现下降。此外，我们证明NAS-NeRF还能实现体积缩小23倍、浮点运算量减少22倍、GPU运行速度快4.7倍的架构，而平均SSIM仅下降5.3%。我们的源代码已在https://saeejithnair.github.io/NAS-NeRF 公开提供。