Research map / 3D vision / NeRF and radiance fields
Mip-Grid: Anti-aliased Grid Representations for Neural Radiance Fields
Seungtae Nam, Daniel Rho, Jong Hwan Ko, Eunbyung Park
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 learned convolutional multi-scale grids from a shared grid | TensoRF · Multi-scale Blender test | average PSNR: 30.70 -> 35.23 | Improved |
| Add learned convolutional multi-scale grids from shared planes | K-Planes · Multi-scale Blender test | average PSNR: 29.76 -> 32.27 | Improved |
| Add learned convolutional multi-scale grids from a shared grid | TensoRF · Multi-scale LLFF test | average PSNR: 26.03 -> 30.48 | Improved |
| Use four independent scale-specific models instead of one shared model | TensoRF · Multi-scale Blender test | average PSNR: 30.70 -> 34.77 | Improved |
| Use four independent scale-specific models instead of one shared model | K-Planes · Multi-scale Blender test | average PSNR: 29.76 -> 32.24 | Improved |
| Use continuous rather than discrete scale coordinates (ablation) | Mip-TensoRF · Multi-scale Blender test, especially unseen resolutions | PSNR: 34.54 -> 34.73 | Improved |
| Add distance from ray origin as a second scale coordinate (ablation) | Mip-TensoRF · Multi-scale Blender test, eight resolutions | average PSNR: 34.54 -> 34.88 | Improved |
| Use fixed Gaussian kernels with increasing standard deviations across scales (ablation) | Mip-TensoRF · Multi-scale Blender test | average PSNR: 30.99 -> 34.05 | Improved |
| Learn kernel weights instead of keeping Gaussian kernels fixed (ablation) | Mip-TensoRF · Multi-scale Blender test | average PSNR: 34.05 -> 35.40 | Improved |
| Increase the number of generated grids from 2 to 4 (ablation) | Mip-TensoRF · Hotdog, multi-scale Blender | average PSNR: 39.43 -> 39.78 | No clear effect |
| Increase kernel size from 3 to 11 (ablation) | Mip-TensoRF · Hotdog, multi-scale Blender, low-resolution rendering | PSNR: 40.49 -> 41.01 | Improved |
| Use kernel sizes that are too small for low-scale grids (ablation) | Mip-TensoRF · Hotdog, multi-scale Blender, low-resolution rendering | PSNR | Hurt |
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.
- Mip-NeRF — Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields (ICCV 2021) · better on 1 of 1 shared result
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
- Zip-NeRF: Anti-Aliased Grid-Based Neural Radiance Fields (ICCV 2023 · similarity 0.94)
- Tri-MipRF: Tri-Mip Representation for Efficient Anti-Aliasing Neural Radiance Fields (ICCV 2023 · similarity 0.92)
- PyNeRF: Pyramidal Neural Radiance Fields (NeurIPS 2023 · similarity 0.92)
- Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields (CVPR 2022 · similarity 0.92)
- Is Vanilla MLP in Neural Radiance Field Enough for Few-shot View Synthesis? (CVPR 2024 · similarity 0.89)
- Rad-NeRF: Ray-decoupled Training of Neural Radiance Field (NeurIPS 2024 · similarity 0.88)
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.
