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

Researchers have applied deep neural networks to image restoration tasks, inwhich they proposed various network architectures, loss functions, and trainingmethods. In particular, adversarial training, which is employed in recentstudies, seems to be a key ingredient to success. In this paper, we show thatsimple convolutional autoencoders (CAEs) built upon only standard networkcomponents, i.e., convolutional layers and skip connections, can outperform thestate-of-the-art methods which employ adversarial training and sophisticatedloss functions. The secret is to employ an evolutionary algorithm toautomatically search for good architectures. Training optimized CAEs byminimizing the $\ell_2$ loss between reconstructed images and their groundtruths using the ADAM optimizer is all we need. Our experimental results showthat this approach achieves 27.8 dB peak signal to noise ratio (PSNR) on theCelebA dataset and 40.4 dB on the SVHN dataset, compared to 22.8 dB and 33.0 dBprovided by the former state-of-the-art methods, respectively.

Exploiting the Potential of Standard Convolutional Autoencoders for Image Restoration by Evolutionary Search