AI Paper Atlas

Research map / 3D vision / NeRF and radiance fields

The Nerfect Match: Exploring NeRF Features for Visual Localization

Qunjie Zhou, Maxim Maximov, Or Litany, Laura Leal-Taixé

ECCV 2024 · 3D vision › NeRF and radiance fields · approach: Dynamic and editable NeRF · 31 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
use NeRF inner features (layers 1-6) instead of positional-encoded 3D points for 2D-3D matching (ablation)NeRFMatch-Mini · Cambridge Landmarkslocalization recall (%): 51.4 -> 59.2Improved
use the last NeRF layer feature f7 instead of middle layers (ablation)NeRFMatch-Mini · Cambridge Landmarksmedian translation (cm): 27.9 -> 61.3Hurt
match against raw 3D coordinates instead of positional encoding (ablation)NeRFMatch-Mini · Cambridge Landmarksmedian translation (cm): 34.3 -> 458.0Hurt
use ImageNet-pretrained ConvFormer backbone instead of from scratch (ablation)NeRFMatch-Mini · Cambridge Landmarksavg recall (%): 50.7 -> 59.2Improved
use coarse-to-fine attention matcher instead of minimal matcher (ablation)NeRFMatch · Cambridge Landmarksavg recall (%): 59.2 -> 71.3Improved
train on multiple scenes instead of per-scene (ablation)NeRFMatch · Cambridge Landmarksavg recall (%): 71.3 -> 65.2Hurt
use iterative refinement instead of photometric optimization refinement (ablation)NeRFMatch (full) · Cambridge Landmarkslatency per step (ms): 398.4 ms -> 141.2 msImproved
use Instant NGP instead of MipNeRF as NeRF (ablation)NeRFMatch · Cambridge Landmarksavg median error (cm/deg)/recall: 13.3/0.3/80.8 -> 28.1/0.5/61.3Hurt
retrieve with NeRF-rendered images instead of real database images (ablation)NeRFMatch · Cambridge Landmarksavg median error (cm): 13.3 -> 14.6Hurt
apply NeRF features for matching on indoor 7-ScenesNeRFMatch · 7-Scenesavg recall (%): 97.8 -> 78.2Hurt

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