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

Images captured in the low-light condition suffer from low visibility andvarious imaging artifacts, e.g., real noise. Existing supervised enlighteningalgorithms require a large set of pixel-aligned training image pairs, which arehard to prepare in practice. Though weakly-supervised or unsupervised methodscan alleviate such challenges without using paired training images, somereal-world artifacts inevitably get falsely amplified because of the lack ofcorresponded supervision. In this paper, instead of using perfectly alignedimages for training, we creatively employ the misaligned real-world images asthe guidance, which are considerably easier to collect. Specifically, wepropose a Cross-Image Disentanglement Network (CIDN) to separately extractcross-image brightness and image-specific content features fromlow/normal-light images. Based on that, CIDN can simultaneously correct thebrightness and suppress image artifacts in the feature domain, which largelyincreases the robustness to the pixel shifts. Furthermore, we collect a newlow-light image enhancement dataset consisting of misaligned training imageswith real-world corruptions. Experimental results show that our model achievesstate-of-the-art performances on both the newly proposed dataset and otherpopular low-light datasets.

Enhancing Low-Light Images in Real World via Cross-Image Disentanglement