Research map / 3D vision / NeRF and radiance fields
Explicit Correspondence Matching for Generalizable Neural Radiance Fields
Yuedong Chen, Haofei Xu, Qianyi Wu, Chuanxia Zheng, Tat‐Jen Cham, Jianfei Cai
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 cross-attention to the CNN feature extractor (ablation) | MatchNeRF · DTU ablation (3 nearest views, foreground) | PSNR: 23.20 -> 26.13 | Improved |
| add self-attention to the CNN feature extractor (ablation) | MatchNeRF · DTU ablation (3 nearest views, foreground) | PSNR: 23.20 -> 23.51 | No clear effect |
| add self-attention on top of cross-attention (ablation) | MatchNeRF · DTU ablation (3 nearest views, foreground) | PSNR: 26.13 -> 26.76 | Improved |
| add ray Transformer to the decoder (ablation) | MatchNeRF · DTU ablation (3 nearest views, foreground) | PSNR: 26.76 -> 26.91 | Improved |
| use group-wise cosine instead of plain cosine similarity (ablation) | MatchNeRF · DTU ablation (3 nearest views, foreground) | PSNR: 26.24 -> 26.91 | Improved |
| use group-wise cosine instead of variance (ablation) | MatchNeRF · DTU ablation (3 nearest views, foreground) | PSNR: 26.36 -> 26.91 | Improved |
| use explicit cosine instead of learned similarity (ablation) | MatchNeRF · DTU ablation (3 nearest views, foreground) | PSNR: 24.87 -> 26.91 | Improved |
| use group-wise cosine instead of feature concatenation (ablation) | MatchNeRF · DTU ablation (3 nearest views, foreground) | PSNR: 24.21 -> 26.91 | Improved |
| increase Transformer blocks from 1 to 6 (ablation) | MatchNeRF · DTU ablation (3 nearest views, foreground) | PSNR: 25.35 -> 26.76 | Improved |
| add 1/4 resolution features to 1/8 features (ablation) | MatchNeRF · DTU ablation (3 nearest views, foreground) | PSNR: 26.09 -> 26.91 | Improved |
| initialize backbone with GMFlow weights instead of random (ablation) | MatchNeRF · DTU (3-view, whole image) | PSNR: 23.44 -> 24.54 | Improved |
| finetune the GMFlow-initialized backbone instead of freezing it (ablation) | MatchNeRF · DTU (3-view, whole image) | PSNR: 20.36 -> 24.54 | Improved |
Closest papers by meaning
- CorresNeRF: Image Correspondence Priors for Neural Radiance Fields (NeurIPS 2023 · similarity 0.91)
- NeRF Analogies: Example-Based Visual Attribute Transfer for NeRFs (CVPR 2024 · similarity 0.89)
- Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fields (CVPR 2022 · similarity 0.89)
- NeRFPrior: Learning Neural Radiance Field as a Prior for Indoor Scene Reconstruction (CVPR 2025 · similarity 0.89)
- Learning Robust Generalizable Radiance Field with Visibility and Feature Augmented Point Representation (ICLR 2024 · similarity 0.89)
- Mask-Based Modeling for Neural Radiance Fields (ICLR 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.
