AI Paper Atlas

Research map / 3D vision / NeRF and radiance fields

UP-NeRF: Unconstrained Pose Prior-Free Neural Radiance Field

Injae Kim, Minhyuk Choi, Hyunwoo J. Kim

NeurIPS 2023 · 3D vision › NeRF and radiance fields · approach: Dynamic and editable NeRF · 2 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
set candidate embedding size to 16 instead of 0 (no embedding) (ablation)UP-NeRF · PhotoTourism meanrotation error (deg): 4.972 -> 1.053Improved
set candidate embedding size to 16 instead of 0 (ablation)UP-NeRF · PhotoTourism meanPSNR: 21.23 -> 23.73Improved
increase candidate embedding size from 16 to 32 (ablation)UP-NeRF · PhotoTourism meanrotation error (deg): 1.053 -> 3.162Hurt
increase candidate embedding size from 8 to 16 (ablation)UP-NeRF · PhotoTourism meanrotation error (deg): 1.286 -> 1.053Improved
increase candidate embedding size from 8 to 16 (ablation)UP-NeRF · PhotoTourism meanSSIM: 0.802 -> 0.800No clear effect
add the candidate head (embedding size 16) (ablation)UP-NeRF · PhotoTourism meantranslation error: 1.042 -> 0.129Improved
add latent appearance modeling and transient head to BARF (BARF-W)BARF · PhotoTourism meanPSNR: 9.13 -> 13.24Improved
add mono-depth loss to BARF-W (BARF-WD)BARF-W · PhotoTourism meanrotation error (deg): 77.50 -> 86.72Mixed
add mono-depth loss to BARF-W (BARF-WD)BARF-W · PhotoTourism meanLPIPS: 0.636 -> 0.557Improved
replace BARF-style baselines with UP-NeRF (feature surrogate, separate transient net, candidate head, depth prior)BARF-WD · PhotoTourism meanPSNR: 12.70 -> 23.73Improved
UP-NeRF without pose prior versus NeRF-W with COLMAP posesUP-NeRF · PhotoTourism meanPSNR: 24.06 -> 23.73No clear effect
remove feature-surrogate bundle adjustment (photometric loss only) (ablation)UP-NeRF · PhotoTourismpose accuracy / renderingImproved

Later papers that compared with this method

1 later accepted paper reports UP-NeRF 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