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 good correspondences is a critical prerequisite in many feature basedtasks. Given a putative correspondence set of an image pair, we propose aneural network which finds correct correspondences by a binary-class classifierand estimates relative pose through classified correspondences. First, weanalyze that due to the imbalance in the number of correct and wrongcorrespondences, the loss function has a great impact on the classificationresults. Thus, we propose a new Guided Loss that can directly use evaluationcriterion (Fn-measure) as guidance to dynamically adjust the objective functionduring training. We theoretically prove that the perfect negative correlationbetween the Guided Loss and Fn-measure, so that the network is always trainedtowards the direction of increasing Fn-measure to maximize it. We then proposea hybrid attention block to extract feature, which integrates the Bayesianattentive context normalization (BACN) and channel-wise attention (CA). BACNcan mine the prior information to better exploit global context and CA cancapture complex channel context to enhance the channel awareness of thenetwork. Finally, based on our Guided Loss and hybrid attention block, acascade network is designed to gradually optimize the result for more superiorperformance. Experiments have shown that our network achieves thestate-of-the-art performance on benchmark datasets. Our code will be availablein https://github.com/wenbingtao/GLHA.

Cascade Network with Guided Loss and Hybrid Attention for Finding Good Correspondences