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/

Zendy Plus

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

Zendy Tools

Zendy Open

Vision Transformers (VTs) are becoming a valuable alternative toConvolutional Neural Networks (CNNs) when it comes to problems involvinghigh-dimensional and spatially organized inputs such as images. However, theirTransfer Learning (TL) properties are not yet well studied, and it is not fullyknown whether these neural architectures can transfer across different domainsas well as CNNs. In this paper we study whether VTs that are pre-trained on thepopular ImageNet dataset learn representations that are transferable to thenon-natural image domain. To do so we consider three well-studied artclassification problems and use them as a surrogate for studying the TLpotential of four popular VTs. Their performance is extensively comparedagainst that of four common CNNs across several TL experiments. Our resultsshow that VTs exhibit strong generalization properties and that these networksare more powerful feature extractors than CNNs.

How Well Do Vision Transformers (VTs) Transfer To The Non-Natural Image Domain? An Empirical Study Involving Art Classification