2023MNRAS.521.6318B

We present an application of auto-encoders to the problem of noise reduction in single-shot astronomical images and explore its suitability for upcoming large-scale surveys. Auto-encoders are a machine learning model that summarizes an input to identify its key features, and then from this knowledge predicts a representation of a different input. The broad aim of our auto-encoder model is to retain morphological information (e.g. non-parametric morphological information) from the survey data while simultaneously reducing the noise contained in the image. We implement an auto-encoder with convolutional and max pooling layers. We test our implementation on images from the Panoramic Survey Telescope and Rapid Response System that contain varying levels of noise and report how successful our auto-encoder is by considering mean squared error, structural similarity index, the second-order moment of the brightest 20 per cent of the galaxy's flux M₂₀, and the Gini coefficient, while noting how the results vary between original images, stacked images, and noise-reduced images. We show that we are able to reduce noise, over many different targets of observations, while retaining the galaxy's morphology, with metric evaluation on a target-by-target analysis. We establish that this process manages to achieve a positive result in a matter of minutes, and by only using one single-shot image compared to multiple survey images found in other noise reduction techniques.