Research map / 3D vision / NeRF and radiance fields
Aug-NeRF: Training Stronger Neural Radiance Fields With Triple-Level Physically-Grounded Augmentations
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Apply triple-level worst-case perturbations to coordinates, intermediate features, and pre-rendering outputs | NeRF · LLFF fern test set | PSNR: 25.17 -> 26.51 | Helped |
| Apply triple-level worst-case perturbations to coordinates, intermediate features, and pre-rendering outputs | NeRF · NeRF-Synthetic ship test set | PSNR: 28.65 -> 29.18 | Hurt |
| Train with triple-level augmentations on noisy supervision | NeRF · Fern + Shot Noise, noise-free test set | PSNR: 15.75 -> 17.00 | Helped |
| Use input-coordinate augmentation only (ablation) | NeRF · LLFF fern test set | PSNR: 25.17 -> 25.30 | No clear effect |
| Use intermediate-feature augmentation only (ablation) | NeRF · LLFF fern test set | PSNR: 25.17 -> 25.39 | Helped |
| Add pre-rendering output augmentation to feature augmentation (ablation) | NeRF with feature augmentation · LLFF fern test set | PSNR: 25.39 -> 26.32 | Helped |
| Replace worst-case perturbations with random augmentation (ablation) | NeRF · LLFF fern test set | PSNR: 26.51 -> 25.28 | Hurt |
| Add explicit l1 sparsity regularization | NeRF · LLFF fern test set | PSNR: 25.17 -> 25.15 | Hurt |
| Add explicit Laplacian regularization | NeRF · LLFF fern test set | PSNR: 25.17 -> 24.89 | Hurt |
| Add explicit total-variation regularization | NeRF · LLFF fern test set | PSNR: 25.17 -> 26.05 | Helped |
| Tune the augmentation strength; overly strong perturbations (e.g., PGD-10) (ablation) | Aug-NeRF · LLFF test set | PSNR | Hurt |
| Apply triple-level augmentation | NeRF · NeRF-Synthetic chair test set | Chamfer Distance: 1.47 × 10^{-2} -> 1.04 × 10^{-2} | Helped |
Closest papers by meaning
- Rad-NeRF: Ray-decoupled Training of Neural Radiance Field (NeurIPS 2024 · similarity 0.93)
- AligNeRF: High-Fidelity Neural Radiance Fields via Alignment-Aware Training (CVPR 2023 · similarity 0.93)
- NeRF-VPT: Learning Novel View Representations with Neural Radiance Fields via View Prompt Tuning (AAAI 2024 · similarity 0.92)
- NeRF On-the-go: Exploiting Uncertainty for Distractor-free NeRFs in the Wild (CVPR 2024 · similarity 0.91)
- Mask-Based Modeling for Neural Radiance Fields (ICLR 2024 · similarity 0.91)
- CF-NeRF: Camera Parameter Free Neural Radiance Fields with Incremental Learning (AAAI 2024 · similarity 0.91)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.