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

Due to the difficulty in collecting large-scale and perfectly aligned pairedtraining data for Under-Display Camera (UDC) image restoration, previousmethods resort to monitor-based image systems or simulation-based methods,sacrificing the realness of the data and introducing domain gaps. In this work,we revisit the classic stereo setup for training data collection -- capturingtwo images of the same scene with one UDC and one standard camera. The key ideais to "copy" details from a high-quality reference image and "paste" them onthe UDC image. While being able to generate real training pairs, this settingis susceptible to spatial misalignment due to perspective and depth of fieldchanges. The problem is further compounded by the large domain discrepancybetween the UDC and normal images, which is unique to UDC restoration. In thispaper, we mitigate the non-trivial domain discrepancy and spatial misalignmentthrough a novel Transformer-based framework that generates well-aligned yethigh-quality target data for the corresponding UDC input. This is made possiblethrough two carefully designed components, namely, the Domain Alignment Module(DAM) and Geometric Alignment Module (GAM), which encourage robust and accuratediscovery of correspondence between the UDC and normal views. Extensiveexperiments show that high-quality and well-aligned pseudo UDC training pairsare beneficial for training a robust restoration network. Code and the datasetare available at https://github.com/jnjaby/AlignFormer.

Generating Aligned Pseudo-Supervision from Non-Aligned Data for Image Restoration in Under-Display Camera