Classification of Star and Galaxy Objects Utilizing Machine Learning Techniques and Deep Neural Networks

Cosmology, a field devoted to comprehending the universe’s vast expanse comprising stars and galaxies, has greatly benefited from advancements in telescopic technology. The heightened resolution capabilities of modern telescopes allow for the capture of intricate celestial images, which, when coupled with machine learning (ML) algorithms, provide a powerful tool for analysis. This research paper delves into the classification of star and galaxy datasets through the utilization of ML techniques, with an emphasis on comparative performance evaluation. Initial findings highlight the effectiveness of the random forest algorithm, yielding an accuracy of 78.91%, surpassing other ML classifiers. To enhance classification precision, a Convolution Neural Network (CNN) model is introduced, yielding an impressive accuracy of 94%. The CNN model’s innate ability to extract key features manifests in its superior classification performance. This work presents a significant stride in star-galaxy classification within the realm of cosmology, employing the amalgamation of advanced telescopic imaging and cutting-edge machine learning techniques.