AI Paper Atlas

Research map / 3D vision / Depth and stereo

Adversarial Exploitation of Data Diversity Improves Visual Localization

Sihang Li, Siqi Tan, Bowen Chang, Jing Zhang, Chen Feng, Yiming Li

ICCV 2025 · 3D vision › Depth and stereo · approach: Scene coordinate regression for localization · 4 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
replace VGG16 with Efficient-B0 feature extractor (ablation)RAP baseline · Cambridge Shoptranslation error (cm): 174 -> 103Improved
add pose augmentation from 3DGS (ablation)RAP · Cambridge Shoptranslation error (cm): 103 -> 75Improved
add appearance augmentation (ablation)RAP · Cambridge Shoptranslation error (cm): 75 -> 60Improved
add ConvNet decoder between features and regression (ablation)RAP · Cambridge Shoptranslation error (cm): 60 -> 52Improved
use Pose Transformer instead of ConvNet decoder (ablation)RAP · Cambridge Shoptranslation error (cm): 52 -> 40Improved
add adversarial discriminator (ablation)RAP · Cambridge Shoptranslation error (cm): 40 -> 33Improved
use our pose augmentation instead of LENS policy (ablation)RAP · Cambridge Landmarkstranslation error (cm): 100 -> 56Improved
add appearance augmentation (ablation)RAP · Aachen Day-Nightrotation error (deg): 75.99 -> 13.70Improved
add augmentation to real training data (100% poses) (ablation)RAP · Cambridge Shopmedian translation error (cm): 98 -> 33Improved
train on 20% real data with augmentation instead of 100% without (ablation)RAP · Cambridge Shopmedian translation error (cm): 98 -> 87Improved
replace NeRF with 3DGS naively in DFNet (ablation)DFNet · Cambridge Landmarksrotation error (deg): 2.90 -> 4.75Hurt
enlarge void zone around test camera to 30 cm / 1.5 deg (ablation)RAP · Cambridge Shopmedian translation error (cm): 33 -> 40Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 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