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

Capturing images with incorrect exposure settings fails to deliver asatisfactory visual experience. Only when the exposure is properly set, can thecolor and details of the images be appropriately preserved. Previous exposurecorrection methods based on convolutions often produce exposure deviation inimages as a consequence of the restricted receptive field of convolutionalkernels. This issue arises because convolutions are not capable of capturinglong-range dependencies in images accurately. To overcome this challenge, wecan apply the Transformer to address the exposure correction problem,leveraging its capability in modeling long-range dependencies to capture globalrepresentation. However, solely relying on the window-based Transformer leadsto visually disturbing blocking artifacts due to the application ofself-attention in small patches. In this paper, we propose a CNN InjectedTransformer (CIT) to harness the individual strengths of CNN and Transformersimultaneously. Specifically, we construct the CIT by utilizing a window-basedTransformer to exploit the long-range interactions among different regions inthe entire image. Within each CIT block, we incorporate a channel attentionblock (CAB) and a half-instance normalization block (HINB) to assist thewindow-based self-attention to acquire the global statistics and refine localfeatures. In addition to the hybrid architecture design for exposurecorrection, we apply a set of carefully formulated loss functions to improvethe spatial coherence and rectify potential color deviations. Extensiveexperiments demonstrate that our image exposure correction method outperformsstate-of-the-art approaches in terms of both quantitative and qualitativemetrics.

CNN Injected Transformer for Image Exposure Correction