Uncovering Tidal Treasures: Automated Classification of Faint Tidal Features in DECaLS Data
Alexander J. Gordon, Annette M. N. Ferguson, Robert G. Mann
arXiv:2404.06487v1 Announce Type: new
Abstract: Tidal features are a key observable prediction of the hierarchical model of galaxy formation and contain a wealth of information about the properties and history of a galaxy. Modern wide-field surveys such as LSST and Euclid will revolutionise the study of tidal features. However, the volume of data will far surpass the capacity to inspect each galaxy to identify the feature visually, thereby motivating an urgent need to develop automated detection methods. This paper presents a visual classification of $sim$2,000 galaxies from the DECaLS survey into different tidal feature categories: arms, streams, shells, and diffuse. Using these labels, we trained a Convolutional Neural Network (CNN) to reproduce the assigned visual classifications. Overall our network performed well and retrieved a median $81.1^{+5.8}_{-6.5}$, $65.7^{+5.0}_{-8.4}$, $91.3^{+6.0}_{-5.9}$, and $82.3^{+1.4}_{-7.9}$ per cent of the actual instances of arm, stream, shell, and diffuse features respectively for just 20 per cent contamination. We verified that the network was classifying the images correctly by using a Gradient-weighted Class Activation Mapping analysis to highlight important regions on the images for a given classification. This is the first demonstration of using CNNs to classify tidal features into sub-categories, and it will pave the way for the identification of different categories of tidal features in the vast samples of galaxies that forthcoming wide-field surveys will deliver.arXiv:2404.06487v1 Announce Type: new
Abstract: Tidal features are a key observable prediction of the hierarchical model of galaxy formation and contain a wealth of information about the properties and history of a galaxy. Modern wide-field surveys such as LSST and Euclid will revolutionise the study of tidal features. However, the volume of data will far surpass the capacity to inspect each galaxy to identify the feature visually, thereby motivating an urgent need to develop automated detection methods. This paper presents a visual classification of $sim$2,000 galaxies from the DECaLS survey into different tidal feature categories: arms, streams, shells, and diffuse. Using these labels, we trained a Convolutional Neural Network (CNN) to reproduce the assigned visual classifications. Overall our network performed well and retrieved a median $81.1^{+5.8}_{-6.5}$, $65.7^{+5.0}_{-8.4}$, $91.3^{+6.0}_{-5.9}$, and $82.3^{+1.4}_{-7.9}$ per cent of the actual instances of arm, stream, shell, and diffuse features respectively for just 20 per cent contamination. We verified that the network was classifying the images correctly by using a Gradient-weighted Class Activation Mapping analysis to highlight important regions on the images for a given classification. This is the first demonstration of using CNNs to classify tidal features into sub-categories, and it will pave the way for the identification of different categories of tidal features in the vast samples of galaxies that forthcoming wide-field surveys will deliver.