Galaxies image classification using artificial bee colony based on orthogonal Gegenbauer moments

Abstract

The development of automated morphological classification schemes for galaxies is important to study the formation and subsequent evolution of galaxies in our universe. This paper proposed a new machine learning method for classifying three types of galaxies image (Hubble types) namely elliptical, lenticulars, and spirals. The proposed approach consists of three stages: In the first stage, the features are extracted from the galaxies image by using Gegenbauer moments that have several properties like scale, rotation, and invariant. However, not all these extract features are relevant; therefore, in the second stage, a swarm algorithm called artificial bee colony (ABC) is used as feature selection method, where ABC has a small number of parameters and its fast convergence to the global solution. The third stage is used to evaluate the performance of the selected features in the classification of galaxies image through using the support vector machine as a classifier. The experimental results are performed based on a sample from EFIGI catalog, and the results show the high performance of the proposed method.