End2End Multi-View Feature Matching with Differentiable Pose Optimization

Barbara Roessle, Matthias Nießner

Publikation: Beitrag in Buch/Bericht/KonferenzbandKonferenzbeitragBegutachtung

Abstract

Erroneous feature matches have severe impact on subsequent camera pose estimation and often require additional, time-costly measures, like RANSAC, for outlier rejection. Our method tackles this challenge by addressing feature matching and pose optimization jointly. To this end, we propose a graph attention network to predict image correspondences along with confidence weights. The resulting matches serve as weighted constraints in a differentiable pose estimation. Training feature matching with gradients from pose optimization naturally learns to downweight outliers and boosts pose estimation on image pairs compared to SuperGlue by 6.7% on ScanNet. At the same time, it reduces the pose estimation time by over 50% and renders RANSAC iterations unnecessary. Moreover, we integrate information from multiple views by spanning the graph across multiple frames to predict the matches all at once. Multi-view matching combined with end-to-end training improves the pose estimation metrics on Matterport3D by 18.5% compared to SuperGlue.

OriginalspracheEnglisch
TitelProceedings - 2023 IEEE/CVF International Conference on Computer Vision, ICCV 2023
Herausgeber (Verlag)Institute of Electrical and Electronics Engineers Inc.
Seiten477-487
Seitenumfang11
ISBN (elektronisch)9798350307184
DOIs
PublikationsstatusVeröffentlicht - 2023
Veranstaltung2023 IEEE/CVF International Conference on Computer Vision, ICCV 2023 - Paris, Frankreich
Dauer: 2 Okt. 20236 Okt. 2023

Publikationsreihe

NameProceedings of the IEEE International Conference on Computer Vision
ISSN (Print)1550-5499

Konferenz

Konferenz2023 IEEE/CVF International Conference on Computer Vision, ICCV 2023
Land/GebietFrankreich
OrtParis
Zeitraum2/10/236/10/23

Fingerprint

Untersuchen Sie die Forschungsthemen von „End2End Multi-View Feature Matching with Differentiable Pose Optimization“. Zusammen bilden sie einen einzigartigen Fingerprint.

Dieses zitieren