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

We present AstroVaDEr, a variational autoencoder designed to performunsupervised clustering and synthetic image generation using astronomicalimaging catalogues. The model is a convolutional neural network that learns toembed images into a low dimensional latent space, and simultaneously optimisesa Gaussian Mixture Model (GMM) on the embedded vectors to cluster the trainingdata. By utilising variational inference, we are able to use the learned GMM asa statistical prior on the latent space to facilitate random sampling andgeneration of synthetic images. We demonstrate AstroVaDEr's capabilities bytraining it on gray-scaled \textit{gri} images from the Sloan Digital SkySurvey, using a sample of galaxies that are classified by Galaxy Zoo 2. Anunsupervised clustering model is found which separates galaxies based onlearned morphological features such as axis ratio, surface brightness profile,orientation and the presence of companions. We use the learned mixture model togenerate synthetic images of galaxies based on the morphological profiles ofthe Gaussian components. AstroVaDEr succeeds in producing a morphologicalclassification scheme from unlabelled data, but unexpectedly places highimportance on the presence of companion objects---demonstrating the importanceof human interpretation. The network is scalable and flexible, allowing forlarger datasets to be classified, or different kinds of imaging data. We alsodemonstrate the generative properties of the model, which allow for realisticsynthetic images of galaxies to be sampled from the learned classificationscheme. These can be used to create synthetic image catalogs or to performimage processing tasks such as deblending.

AstroVaDEr: Astronomical Variational Deep Embedder for Unsupervised Morphological Classification of Galaxies and Synthetic Image Generation