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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Deep generative models (DGMs) have the potential to revolutionize diagnosticimaging. Generative adversarial networks (GANs) are one kind of DGM which arewidely employed. The overarching problem with deploying GANs, and other DGMs,in any application that requires domain expertise in order to actually use thegenerated images is that there generally is not adequate or automatic means ofassessing the domain-relevant quality of generated images. In this work, wedemonstrate several objective tests of images output by two popular GANarchitectures. We designed several stochastic context models (SCMs) of distinctimage features that can be recovered after generation by a trained GAN. Severalof these features are high-order, algorithmic pixel-arrangement rules which arenot readily expressed in covariance matrices. We designed and validatedstatistical classifiers to detect specific effects of the known arrangementrules. We then tested the rates at which two different GANs correctlyreproduced the feature context under a variety of training scenarios, anddegrees of feature-class similarity. We found that ensembles of generatedimages can appear largely accurate visually, and show high accuracy in ensemblemeasures, while not exhibiting the known spatial arrangements. Furthermore,GANs trained on a spectrum of distinct spatial orders did not respect the givenprevalence of those orders in the training data. The main conclusion is thatSCMs can be engineered to quantify numerous errors, per image, that may not becaptured in ensemble statistics but plausibly can affect subsequent use of theGAN-generated images.

A Method for Evaluating Deep Generative Models of Images via Assessing the Reproduction of High-order Spatial Context