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

Blind image super-resolution(SR) is a long-standing task in CV that aims torestore low-resolution images suffering from unknown and complex distortions.Recent work has largely focused on adopting more complicated degradation modelsto emulate real-world degradations. The resulting models have madebreakthroughs in perceptual loss and yield perceptually convincing results.However, the limitation brought by current generative adversarial networkstructures is still significant: treating pixels equally leads to the ignoranceof the image's structural features, and results in performance drawbacks suchas twisted lines and background over-sharpening or blurring. In this paper, wepresent A-ESRGAN, a GAN model for blind SR tasks featuring an attention U-Netbased, multi-scale discriminator that can be seamlessly integrated with othergenerators. To our knowledge, this is the first work to introduce attentionU-Net structure as the discriminator of GAN to solve blind SR problems. And thepaper also gives an interpretation for the mechanism behind multi-scaleattention U-Net that brings performance breakthrough to the model. Throughcomparison experiments with prior works, our model presents state-of-the-artlevel performance on the non-reference natural image quality evaluator metric.And our ablation studies have shown that with our discriminator, the RRDB basedgenerator can leverage the structural features of an image in multiple scales,and consequently yields more perceptually realistic high-resolution imagescompared to prior works.

A-ESRGAN: Training Real-World Blind Super-Resolution with Attention U-Net Discriminators