CARAD: Computer-aided Analysis of Radio Astronomy Data

In his thesis, Steven Machetho introduces a new method to process huge amounts of radio telescope data using intelligent, efficient and easily interpretable algorithms.

Machetho developed a light and transparent computer vision approach that is able to recognize and classify the shapes of radio galaxies, retrieve similar cosmic structures from large archives, and detect rare anomalies – all with significantly less computing power than traditional deep-learning models require.

As the next generation of telescopes, such as the Square Kilometre Array, begins to produce unprecedented amounts of data, this approach offers timely and essential advances for the field of astronomy.

Dissertation: https://hdl.handle.net/11370/4966659b-43bb-40f1-a474-ecea49357713