Deep Learning-based Imaging in Radio Interferometry

TL;DR

This paper introduces a deep learning approach using convolutional neural networks to efficiently and reproducibly reconstruct radio interferometric images from incomplete Fourier data, improving over traditional methods.

Contribution

The paper presents a novel CNN-based method inspired by super-resolution models for radio image reconstruction, demonstrating comparable or improved performance with greater efficiency.

Findings

Source angles and sizes are accurately reconstructed.

Flux recovery shows some scatter but is comparable to existing methods.

The method is faster and more reproducible than traditional techniques.

Abstract

The sparse layouts of radio interferometers result in an incomplete sampling of the sky in Fourier space which leads to artifacts in the reconstructed images. Cleaning these systematic effects is essential for the scientific use of radiointerferometric images. Established reconstruction methods are often time-consuming, require expert-knowledge, and suffer from a lack of reproducibility. We have developed a prototype Deep Learning-based method that generates reproducible images in an expedient fashion. To this end, we take advantage of the efficiency of Convolutional Neural Networks to reconstruct image data from incomplete information in Fourier space. The Neural Network architecture is inspired by super-resolution models that utilize residual blocks. Using simulated data of radio galaxies that are composed of Gaussian components we train Deep Learning models whose reconstruction capability is quantified using various measures. The reconstruction performance is evaluated on clean and noisy input data by comparing the resulting predictions with the true source images. We find that source angles and sizes are well reproduced, while the recovered fluxes show substantial scatter, albeit not worse than existing methods without fine-tuning. Finally, we propose more advanced approaches using Deep Learning that include uncertainty estimates and a concept to analyze larger images.