AI Paper Atlas

Research map / Point clouds / Point cloud classification

Unified and Efficient Point-Line Local Features

François Costa, Raphael Kreft, Felix Möller, Hardik Shah, Ramanathan Rajaraman, Eckhard Goedeke, Shaohui Liu, Rémi Pautrat, Marc Pollefeys

ECCV 2026 · Point clouds › Point cloud classification · approach: Point cloud detection and tracking · 0 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 joint line training to the point network (ablation)UPAL · MegaDepth relative poseAUC @5deg: 55.2 -> 58.2Improved
remove angle field branch from DeepLSD, use gradient angle (ablation)DeepLSD · HPatches line detectionrepeatability / loc error: 27.2 -> 28.2 repeatability; 1.432 -> 1.341 loc errImproved
compute gradient angle in parallel on GPU (ablation)fast LSD · HPatches line detectionlatency (ms): 36 -> 27Improved
downsample seed grid with stride 2 (ablation)fast LSD · HPatches line detectionlatency (ms): 27 -> 19Improved
keep only a subset of seed points (ablation)fast LSD · HPatches line detectionlatency (ms): 19 -> 17Improved
offload image preprocessing to GPU (ablation)fast LSD · HPatches line detectionlatency (ms): 17 -> 11Improved
add lines to points for visual localizationUPAL · 7Scenes Stairsaccuracy: 49.1 -> 54.6Improved
use UPAL point features vs ALIKEDUPAL · MegaDepth relative poseAUC @5deg: 59.8 -> 58.2Hurt
use UPAL vs XFeat on indoor ScanNetUPAL · ScanNet relative poseAUC @20deg: 45.7 -> 41.4Hurt
unified network vs ALIKED + DeepLSD pipelineUPAL · Runtime (RTX 2080 Ti)GPU latency (ms): 286 -> 70Improved
UPAL vs DISK on ETH3D triangulationUPAL · ETH3D triangulationaccuracy @1cm (%): 69.22 -> 64.80Hurt
UPAL vs DeepLSD on RDNIM homography estimationUPAL · RDNIMH estimation @5px: 60.5 -> 58.0Hurt

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