AI Paper Atlas

Research map / 3D vision / NeRF and radiance fields

GeCoNeRF: Few-shot Neural Radiance Fields via Geometric Consistency

Min-Seop Kwak, Jiuhn Song, Seungryong Kim

ICML 2023 · 3D vision › NeRF and radiance fields · approach: Sparse-input NeRF regularization · 8 citations (OpenAlex)

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 changedWhereMeasuredEffect
Add depth-guided feature-level consistency between rendered novel views and warped input images (ablation)mip-NeRF · LLFF (3-view)PSNR: 14.62 -> 18.10Improved
Add an occlusion mask to exclude geometrically inconsistent warped regions (ablation)GeCoNeRF with feature-level consistency · LLFF (3-view)PSNR: 18.10 -> 18.24Improved
Add progressive training (ablation)GeCoNeRF with consistency and occlusion mask · LLFF (3-view)PSNR: 18.24 -> 18.46Improved
Add edge-aware disparity regularization (ablation)GeCoNeRF · LLFF (3-view)PSNR: 18.46 -> 18.55Improved
Remove progressive positional-encoding annealing (ablation)GeCoNeRF · NeRF-Synthetic (3-view)PSNR: 19.23 -> 18.50Hurt
Replace progressively growing pose-angle intervals with large intervals from the start (ablation)GeCoNeRF · NeRF-Synthetic (3-view)PSNR: 19.23 -> 16.96Hurt
Remove both progressive positional-encoding annealing and progressive pose sampling (ablation)GeCoNeRF · NeRF-Synthetic (3-view)PSNR: 19.23 -> 17.04Hurt
Use feature-level consistency instead of masked pixel-level photometric consistency (ablation)GeCoNeRF · LLFF (3-view)PSNR: 17.98 -> 18.55Improved
Use GeCoNeRF rather than InfoNeRFFew-shot NeRF · NeRF-Synthetic (3-view)PSNR: 18.65 -> 19.23Improved
Use GeCoNeRF rather than RegNeRFFew-shot NeRF · LLFF (3-view)PSNR: 19.08 -> 18.77Hurt
Use GeCoNeRF rather than RegNeRFFew-shot NeRF · LLFF (3-view)SSIM: 0.587 -> 0.596Improved
Use GeCoNeRF rather than RegNeRFFew-shot NeRF · LLFF (3-view)LPIPS: 0.336 -> 0.338Hurt

Later papers that compared with this method

1 later accepted paper reports GeCoNeRF as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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.

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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy