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

Image super-resolution (SR) is a technique to recover lost high-frequencyinformation in low-resolution (LR) images. Spatial-domain information has beenwidely exploited to implement image SR, so a new trend is to involvefrequency-domain information in SR tasks. Besides, image SR is typicallyapplication-oriented and various computer vision tasks call for image arbitrarymagnification. Therefore, in this paper, we study image features in thefrequency domain to design a novel scale-arbitrary image SR network. First, westatistically analyze LR-HR image pairs of several datasets under differentscale factors and find that the high-frequency spectra of different imagesunder different scale factors suffer from different degrees of degradation, butthe valid low-frequency spectra tend to be retained within a certaindistribution range. Then, based on this finding, we devise an adaptivescale-aware feature division mechanism using deep reinforcement learning, whichcan accurately and adaptively divide the frequency spectrum into thelow-frequency part to be retained and the high-frequency one to be recovered.Finally, we design a scale-aware feature recovery module to capture and fusemulti-level features for reconstructing the high-frequency spectrum atarbitrary scale factors. Extensive experiments on public datasets show thesuperiority of our method compared with state-of-the-art methods.

A Scale-Arbitrary Image Super-Resolution Network Using Frequency-domain Information