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

Image-to-image translation has recently achieved remarkable results. Butdespite current success, it suffers from inferior performance when translationsbetween classes require large shape changes. We attribute this to thehigh-resolution bottlenecks which are used by current state-of-the-artimage-to-image methods. Therefore, in this work, we propose a novel deephierarchical Image-to-Image Translation method, called DeepI2I. We learn amodel by leveraging hierarchical features: (a) structural information containedin the shallow layers and (b) semantic information extracted from the deeplayers. To enable the training of deep I2I models on small datasets, we proposea novel transfer learning method, that transfers knowledge from pre-trainedGANs. Specifically, we leverage the discriminator of a pre-trained GANs (i.e.BigGAN or StyleGAN) to initialize both the encoder and the discriminator andthe pre-trained generator to initialize the generator of our model. Applyingknowledge transfer leads to an alignment problem between the encoder andgenerator. We introduce an adaptor network to address this. On many-classimage-to-image translation on three datasets (Animal faces, Birds, and Foods)we decrease mFID by at least 35% when compared to the state-of-the-art.Furthermore, we qualitatively and quantitatively demonstrate that transferlearning significantly improves the performance of I2I systems, especially forsmall datasets. Finally, we are the first to perform I2I translations fordomains with over 100 classes.

DeepI2I: Enabling Deep Hierarchical Image-to-Image Translation by Transferring from GANs