AI Paper Atlas

Research map / Detection and segmentation / CNN features

RDD: Improving Structure from Motion with a Robust Feature Detector and Descriptor

G Chen, Tianwen Fu, Haiwei Chen, Wenbin Teng, Hanyuan Xiao, Yajie Zhao

IJCV 2026 · Detection and segmentation › CNN features · approach: Context for hard detection conditions · 1 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 deformable-transformer descriptor branch (RDD) instead of best prior detector (DeDoDe-V2-G), MNN matchingRDD · MegaDepth-ViewAUC@5deg: 33.1 -> 38.3Improved
use RDD instead of DeDoDe-V2-G on Air-to-Ground pairssparse MNN matching · Air-to-Ground benchmarkAUC@5deg: 31.5 -> 41.4Improved
use larger patch size s=8 (ablation)RDD · MegaDepth-1500AUC@5deg: 48.2 -> 44.6Hurt
use fewer sampled points per query (N=4) in deformable attention (ablation)RDD · MegaDepth-1500AUC@5deg: 48.2 -> 46.5Hurt
add the dedicated keypoint branch (ablation)RDD · MegaDepth-1500AUC@5deg: 44.1 -> 48.2Improved
train the two branches separately instead of jointly (ablation)RDD · MegaDepth-1500AUC@5deg: 42.8 -> 48.2Improved
train the two branches separately instead of jointly (ablation)RDD* · MegaDepth-1500AUC@5deg: 46.9 -> 51.3Improved
add epipolar match refinement to semi-dense matching (ablation)RDD* · MegaDepth-1500AUC@5deg: 41.3 -> 51.3Improved
add Air-to-Ground data to training (ablation)RDD · MegaDepth-1500AUC@5deg: 47.4 -> 48.2Improved
add Air-to-Ground data to training (ablation)RDD* · MegaDepth-1500AUC@5deg: 50.4 -> 51.3Improved
use larger patch size s=8 (ablation)RDD* · MegaDepth-1500AUC@5deg: 51.3 -> 49.4Hurt
use LightGlue matcher instead of MNNRDD · MegaDepth-ViewAUC@5deg: 38.3 -> 54.2Improved

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