AI Paper Atlas

Research map / 3D vision / NeRF and radiance fields

NerfingMVS: Guided Optimization of Neural Radiance Fields for Indoor Multi-View Stereo

Yi Wei, Shaohui Liu, Yongming Rao, Wang Zhao, Jiwen Lu, Jie Zhou

ICCV 2021 · 3D vision › NeRF and radiance fields · approach: Dynamic and editable NeRF · 273 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 adapted depth priors to guide NeRF sampling (ablation)NeRF · ScanNet scene0521Abs Rel: 0.302 -> 0.053Improved
add confidence-based filter (ablation)NeRF + depth priors · ScanNet scene0521Abs Rel: 0.053 -> 0.051Improved
add confidence filter to full method (ablation)Ours · ScanNet (8 scenes)Abs Rel: 0.0635 -> 0.0614Improved
add NeRF optimization over the depth priors (ablation)depth priors · ScanNet scene0521Abs Rel: 0.067 -> 0.053Improved
add adaptive sampling range (ablation)guided optimization · ScanNet scene0521Abs Rel: 0.065 -> 0.056Improved
add bounds alpha_l and alpha_h (ablation)guided optimization · ScanNet scene0521Abs Rel: 0.056 -> 0.051Improved
apply CVD optimization on adapted priorsadapted depth priors · ScanNet (8 scenes)Abs Rel: 0.0844 -> 0.0886Hurt
apply guided NeRF optimization on adapted priorsadapted depth priors · ScanNet (8 scenes)Abs Rel: 0.0844 -> 0.0635Improved
finetune monocular depth net on scene COLMAP reconstructionMannequin · ScanNet (8 scenes)Abs Rel: 0.1554 -> 0.0844Improved
use guided optimization instead of vanilla NeRFNeRF · ScanNet seen viewsPSNR: 28.62 -> 31.55Improved
use guided optimization instead of vanilla NeRFNeRF · ScanNet novel viewsPSNR: 15.76 -> 18.07Improved
set K=4 instead of K=2 (ablation)Ours · ScanNet scene0521Abs Rel: 0.055 -> 0.051Improved

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