AI Paper Atlas

Research map / 3D vision / 3D generation

GeoGraph: Graph-based multi-view object detection with geometric cues end-to-end

Ahmed Samy Nassar, Stefano D’Aronco, Sébastien Lefèvre, Jan D. Wegner

ECCV 2020 · 3D vision › 3D generation · approach: Learning 3D reconstruction from images · 28 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 SSD with EfficientDet detectorSSD-ReID-Geo · Pasadena Multi-View ReIDDetection mAP: 0.682 -> 0.742Improved
replace SSD with EfficientDet detectorSSD-ReID-Geo · Mapillary traffic signsDetection mAP: 0.902 -> 0.919Improved
use GNN instead of siamese CNN (ablation)Siamese CNN · Pasadena Multi-View ReID (GT crops)F1-score: 0.509 -> 0.601Improved
use GNN instead of siamese CNN (ablation)Siamese CNN · Mapillary traffic signs (GT crops)F1-score: 0.721 -> 0.823Improved
add geometric camera cues (GNN-Geo) (ablation)GNN · Pasadena Multi-View ReID (GT crops)F1-score: 0.601 -> 0.640Improved
add geometric camera cues (GNN-Geo) (ablation)GNN · Mapillary traffic signs (GT crops)F1-score: 0.823 -> 0.873Improved
increase views from 2 to 4 (ablation)GeoGraph · Pasadena Multi-View ReIDGeo-localization error (m): 2.94 -> 2.75Improved
increase views from 2 to 4 (ablation)GeoGraph · Pasadena Multi-View ReIDRe-ID mAP: 0.754 -> 0.763Improved
increase views from 2 to 6 (ablation)GeoGraph · Mapillary traffic signsRe-ID mAP: 0.902 -> 0.924Improved
increase views from 2 to 6 (ablation)GeoGraph · Mapillary traffic signsGeo-localization error (m): 3.88 -> 4.21Hurt
fully connected dense edges including within-image nodes (ablation)GNN graph construction · Pasadena Multi-View ReID and Mapillary traffic signsRe-IDImproved

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