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

In real applications, object detectors based on deep networks still facechallenges of the large domain gap between the labeled training data andunlabeled testing data. To reduce the gap, recent techniques are proposed byaligning the image/instance-level features between source and unlabeled targetdomains. However, these methods suffer from the suboptimal problem mainlybecause of ignoring the category information of object instances. To tacklethis issue, we develop a fine-grained domain alignment approach with awell-designed domain classifier bank that achieves the instance-level alignmentrespecting to their categories. Specifically, we first employ the mean teacherparadigm to generate pseudo labels for unlabeled samples. Then we implement theclass-level domain classifiers and group them together, called domainclassifier bank, in which each domain classifier is responsible for aligningfeatures of a specific class. We assemble the bare object detector with theproposed fine-grained domain alignment mechanism as the adaptive detector, andoptimize it with a developed crossed adaptive weighting mechanism. Extensiveexperiments on three popular transferring benchmarks demonstrate theeffectiveness of our method and achieve the new remarkable state-of-the-arts.

Learning a Domain Classifier Bank for Unsupervised Adaptive Object Detection