Automatic detection and quantitative assessment of peculiar galaxy pairs in Sloan Digital Sky Survey

Abstract

We applied computational tools for automatic detection of peculiar galaxy pairs. We first detected in SDSS DR7 ∼400,000 galaxy images with i magnitude <18 that had more than one point spread function, and then applied a machine learning algorithm that detected ∼26,000 galaxy images that had morphology similar to the morphology of galaxy mergers. That dataset was mined using a novelty detection algorithm, producing a short list of 500 most peculiar galaxies as quantitatively determined by the algorithm. Manual examination of these galaxies showed that while most of the galaxy pairs in the list were not necessarily peculiar, numerous unusual galaxy pairs were detected. In this paper we describe the protocol and computational tools used for the detection of peculiar mergers, and provide examples of peculiar galaxy pairs that were detected.

An automatic taxonomy of galaxy morphology using unsupervised machine learning

“…Shamir, Holincheck & Wallin (2013) used a pre-defined training set with supervised and unsupervised algorithms to classify galaxy mergers. Shamir & Wallin (2014) combined supervised and unsupervised techniques in an outlier technique to identify peculiar galaxy pairs in 400 000 SDSS images. Existing works incorporating unsupervised algorithms to classify images of galaxies all use the collation of a training data set by pre-labelling galaxies.…”

Morphology-based Query for Galaxy Image Databases

“…However, the bandwidth of manual annotation cannot satisfy the data collection capacity of the current digital sky surveys such as the Dark Energy Survey (DES), and its reliance on the processing power of the human brain limits the opportuni-ties to improve its bandwidth, making it even more difficult to effectively analyze future sky surveys such as LSST. That reinforces the use of automation to generate catalogs of galaxy morphology (Shamir 2009;Dieleman et al 2015), and automatically generated catalogs have been collected and published (Huertas-Company et al 2010;Fasano et al 2012;Shamir and Wallin 2014;Gravet et al 2015;Kuminski and Shamir 2016;Huertas-Company et al 2016).…”

“…The Wndchrm feature set also contains color descriptors (Shamir and Tarakhovsky 2012;Shamir 2012b;Shamir et al 2010). However, these color features have shown mild contribution to the task of galaxy image analysis (Shamir 2009;Kuminski et al 2014;Shamir and Wallin 2014), and are therefore not used in this experiment. All images are treated as grayscale images.…”

“…However, automatic annotation of galaxies by their morphology is merely one task related to galaxy image analysis that can be performed by computers. Other tasks can include automatic detection of peculiar galaxies in large datasets (Shamir 2012a;Shamir and Wallin 2014), or grouping galaxies by their visual similarities using unsupervised machine learning (Shamir et al 2013;Schutter and Shamir 2015).…”

A Catalog of Automatically Detected Ring Galaxy Candidates in PanSTARSS

“…Due to the large sizes of these databases, effective identification of these objects requires automation, leading to the development of automatic methods of identifying peculiar objects in large databases of galaxy images (Shamir 2012(Shamir , 2016Shamir & Wallin 2014). Here we describe an automatic image analysis method that can identify ring galaxies and apply the method to mine through ∼3×10 6 galaxies imaged by the Panoramic Survey Telescope and Rapid Response System (Hodapp et al 2004;Chambers et al 2016;Flewelling et al 2016) to compile a catalog of ring galaxy candidates.…”