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

Despite the remarkable achievement of recent underwater image restorationtechniques, the lack of labeled data has become a major hurdle for furtherprogress. In this work, we propose a mean-teacher based Semi-supervisedUnderwater Image Restoration (Semi-UIR) framework to incorporate the unlabeleddata into network training. However, the naive mean-teacher method suffers fromtwo main problems: (1) The consistency loss used in training might becomeineffective when the teacher's prediction is wrong. (2) Using L1 distance maycause the network to overfit wrong labels, resulting in confirmation bias. Toaddress the above problems, we first introduce a reliable bank to store the"best-ever" outputs as pseudo ground truth. To assess the quality of outputs,we conduct an empirical analysis based on the monotonicity property to selectthe most trustworthy NR-IQA method. Besides, in view of the confirmation biasproblem, we incorporate contrastive regularization to prevent the overfittingon wrong labels. Experimental results on both full-reference and non-referenceunderwater benchmarks demonstrate that our algorithm has obvious improvementover SOTA methods quantitatively and qualitatively. Code has been released athttps://github.com/Huang-ShiRui/Semi-UIR.

Contrastive Semi-supervised Learning for Underwater Image Restoration via Reliable Bank