Ganalyzer: A tool for automatic galaxy image analysis

TL;DR

Ganalyzer is an efficient, model-based tool for automatic galaxy image analysis that measures spirality to classify galaxies, outperforming manual methods in speed and accuracy for large datasets.

Contribution

It introduces a novel, spiral-based analysis method that is simple, fast, and more accurate than manual classification, suitable for large-scale galaxy surveys.

Findings

Analyzes ~10 million galaxy images in five days on a standard desktop.

Provides accurate galaxy classification based on spirality measurement.

Available as free software for large dataset analysis.

Abstract

We describe Ganalyzer, a model-based tool that can automatically analyze and classify galaxy images. Ganalyzer works by separating the galaxy pixels from the background pixels, finding the center and radius of the galaxy, generating the radial intensity plot, and then computing the slopes of the peaks detected in the radial intensity plot to measure the spirality of the galaxy and determine its morphological class. Unlike algorithms that are based on machine learning, Ganalyzer is based on measuring the spirality of the galaxy, a task that is difficult to perform manually, and in many cases can provide a more accurate analysis compared to manual observation. Ganalyzer is simple to use, and can be easily embedded into other image analysis applications. Another advantage is its speed, which allows it to analyze ~10,000,000 galaxy images in five days using a standard modern desktop computer. These capabilities can make Ganalyzer a useful tool in analyzing large datasets of galaxy images collected by autonomous sky surveys such as SDSS, LSST or DES. The software is available for free download at http://vfacstaff.ltu.edu/lshamir/downloads/ganalyzer, and the data used in the experiment are available at http://vfacstaff.ltu.edu/lshamir/downloads/ganalyzer/GalaxyImages.zip.