Morphology and clustering of galaxies images using transfer learning and unsupervised machine learning

Clustering galaxies based on structural properties reveals important details about their evolution and development. The morphology of a galaxy reveals a lot about its dynamic and merger history. It is closely linked to a number of physical parameters, including mass, star formation background, and mass distribution. Unsupervised machine learning, does not include labeled data, is appropriate for morphological analysis of recent and upcoming surveys, with its approach based on extensive computer vision techniques that measure visual similarities between various forms of morphology. The proposed method perceives similarities between collections of galaxy images automatically. This article makes use of the K-means clustering and threshold functions. The whole dataset is partitioned into K-clusters using the K- mean algorithm.