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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Template matching is a fundamental task in computer vision and has beenstudied for decades. It plays an essential role in manufacturing industry forestimating the poses of different parts, facilitating downstream tasks such asrobotic grasping. Existing methods fail when the template and source imageshave different modalities, cluttered backgrounds or weak textures. They alsorarely consider geometric transformations via homographies, which commonlyexist even for planar industrial parts. To tackle the challenges, we propose anaccurate template matching method based on differentiable coarse-to-finecorrespondence refinement. We use an edge-aware module to overcome the domaingap between the mask template and the grayscale image, allowing robustmatching. An initial warp is estimated using coarse correspondences based onnovel structure-aware information provided by transformers. This initialalignment is passed to a refinement network using references and aligned imagesto obtain sub-pixel level correspondences which are used to give the finalgeometric transformation. Extensive evaluation shows that our method issignificantly better than state-of-the-art methods and baselines, providinggood generalization ability and visually plausible results even on unseen realdata.

Learning Accurate Template Matching with Differentiable Coarse-to-Fine Correspondence Refinement