Research map / 3D vision / NeRF and radiance fields
From NeRFLiX to NeRFLiX++: A General NeRF-Agnostic Restorer Paradigm
Kun Zhou, Wenbo Li, Nianjuan Jiang, Xiaoguang Han, Jiangbo Lu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add WKS supervision instead of L1 (ablation) | G-IVM · LLFF-P1 | PSNR: 27.03 -> 27.38 | Improved |
| increase top-K patches from K=1 to K=5 (ablation) | WKS · LLFF-P1 | PSNR: 27.21 -> 27.38 | Improved |
| increase K from 5 to 20 (ablation) | WKS · LLFF-P1 | PSNR: 27.38 -> 27.37 | No clear effect |
| add optical flow guidance (ablation) | G-IVM · LLFF-P1 | PSNR: 27.21 -> 27.38 | Improved |
| align to aggregated previous-level feature instead of F^l (ablation) | G-IVM · LLFF-P1 | PSNR: 26.97 -> 27.38 | Improved |
| add pyramid aggregation and reconstruction levels (ablation) | G-IVM · LLFF-P1 | PSNR: 26.84 -> 27.38 | Improved |
| estimate flow at 1/4 resolution rather than full (ablation) | G-IVM · LLFF-P1 | PSNR: 27.25 -> 27.38 | Improved |
| estimate flow at 1/16 instead of 1/4 (ablation) | G-IVM · LLFF-P1 | PSNR: 27.38 -> 27.17 | Hurt |
| add deep generative simulator to hand-crafted degradations (ablation) | NeRFLiX++ simulator · LLFF-P1 | PSNR: 27.14 -> 27.38 | Improved |
| train on NDS-P degradation instead of BSR (ablation) | EDVR · LLFF-P1 | PSNR: 26.01 -> 26.98 | Improved |
| replace NeRFLiX with NeRFLiX++ guided mixer | TensoRF · Noisy LLFF Synthetic | PSNR: 24.12 -> 25.39 | Improved |
| replace NeRFLiX with NeRFLiX++ | TensoRF · LLFF-P1 | PSNR: 27.39 -> 27.38 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Plenoxels — Plenoxels: Radiance Fields Without Neural Networks (CVPR 2022) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- NeRFLix: High-Quality Neural View Synthesis by Learning a Degradation-Driven Inter-Viewpoint MiXer (CVPR 2023 · similarity 0.98)
- NeRF-VPT: Learning Novel View Representations with Neural Radiance Fields via View Prompt Tuning (AAAI 2024 · similarity 0.92)
- Optimize the Unseen - Fast NeRF Cleanup with Free Space Prior (NeurIPS 2025 · similarity 0.91)
- NeRF On-the-go: Exploiting Uncertainty for Distractor-free NeRFs in the Wild (CVPR 2024 · similarity 0.91)
- Rad-NeRF: Ray-decoupled Training of Neural Radiance Field (NeurIPS 2024 · similarity 0.90)
- LuSh-NeRF: Lighting up and Sharpening NeRFs for Low-light Scenes (NeurIPS 2024 · similarity 0.90)
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.
