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

The convolutional neural network (CNN) remains an essential tool in solvingcomputer vision problems. Standard convolutional architectures consist ofstacked layers of operations that progressively downscale the image. Aliasingis a well-known side-effect of downsampling that may take place: it causeshigh-frequency components of the original signal to become indistinguishablefrom its low-frequency components. While downsampling takes place in themax-pooling layers or in the strided-convolutions in these models, there is noexplicit mechanism that prevents aliasing from taking place in these layers.Due to the impressive performance of these models, it is natural to suspectthat they, somehow, implicitly deal with this distortion. The question we aimto answer in this paper is simply: "how and to what extent do CNNs counteractaliasing?" We explore the question by means of two examples: In the first, weassess the CNNs capability of distinguishing oscillations at the input, showingthat the redundancies in the intermediate channels play an important role insucceeding at the task; In the second, we show that an image classifier CNNwhile, in principle, capable of implementing anti-aliasing filters, does notprevent aliasing from taking place in the intermediate layers.

How Convolutional Neural Networks Deal with Aliasing