降低计算与数据需求，让NeRF更快更省地生成新视角图像。
Low-Cost Neural Radiance Fields
- 在低视角数下改进TensoRF，加入深度监督提升效果。
- 通过简化网络结构和降采样输入，显著降低训练时间与资源消耗。
- 提出多种新架构变体，为轻量级NeRF设计提供参考方向。
Neural Radiance Fields (NeRF) 虽能实现高质量的新视角合成，但其长训练时间和对密集输入视角的依赖限制了普及。本文对比研究了三种加速版NeRF（DS-NeRF、TensoRF、HashNeRF），并探索面向低算力、低数据场景的改进方案。首先，在TensoRF中引入基于COLMAP关键点的深度监督损失（TensoRF-DS），在LLFF数据集上于少视角条件下评估；其次，消融TensoRF的特征解码MLP，并研究输入降采样对合成玩具场景（Lego）的PSNR与运行时的影响；第三，提出四种HashNeRF颜色与密度网络的架构变体（含残差与卷积设计），在相同迭代预算下报告PSNR与训练时间的权衡。在等时评估下，各项改进未全面超越已有基线，但明确了哪些方法适用于受限环境，并揭示了未来设计的关键问题。
原文摘要 · Abstract (English)
Neural Radiance Fields (NeRF) achieve high-quality novel-view synthesis, but their long training times and reliance on dense input views limit accessibility. We present a comparative study of three accelerated NeRF variants - DS-NeRF, TensoRF, and HashNeRF and explore extensions targeted at the low-compute, low-data regime. First, we add a depth-supervision loss derived from COLMAP keypoints to TensoRF (TensoRF-DS) and evaluate it on the LLFF dataset under reduced view counts. Second, we ablate the feature-decoding MLP of TensoRF and study the effect of input downsampling on PSNR and runtime on the synthetic Lego scene. Third, we propose four architectural variants of the HashNeRF color and density networks, including residual and convolutional designs, and report PSNR/training-time tradeoffs under matched iteration budgets. Under iso-time evaluation, none of our extensions conclusively outperform the published baselines, but the experiments characterize which extensions transfer to constrained settings and surface design questions for future work.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。