English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Finding localized correspondences across different images of the same objectis crucial to understand its geometry. In recent years, this problem has seenremarkable progress with the advent of deep learning-based local image featuresand learnable matchers. Still, learnable matchers often underperform when thereexists only small regions of co-visibility between image pairs (i.e. widecamera baselines). To address this problem, we leverage recent progress incoarse single-view geometry estimation methods. We propose LFM-3D, a LearnableFeature Matching framework that uses models based on graph neural networks andenhances their capabilities by integrating noisy, estimated 3D signals to boostcorrespondence estimation. When integrating 3D signals into the matcher model,we show that a suitable positional encoding is critical to effectively make useof the low-dimensional 3D information. We experiment with two different 3Dsignals - normalized object coordinates and monocular depth estimates - andevaluate our method on large-scale (synthetic and real) datasets containingobject-centric image pairs across wide baselines. We observe strong featurematching improvements compared to 2D-only methods, with up to +6% total recalland +28% precision at fixed recall. Additionally, we demonstrate that theresulting improved correspondences lead to much higher relative posing accuracyfor in-the-wild image pairs - up to 8.6% compared to the 2D-only approach.

LFM-3D: Learnable Feature Matching Across Wide Baselines Using 3D Signals