NORMA eResearch @NCI Library

Deep Learning Techniques for Astronomical Object Classification

Dalvi, Yogiraj Subhash (2022) Deep Learning Techniques for Astronomical Object Classification. Masters thesis, Dublin, National College of Ireland.

[thumbnail of Master of Science]
PDF (Master of Science)
Download (1MB) | Preview
[thumbnail of Configuration manual]
PDF (Configuration manual)
Download (2MB) | Preview


As there are an infinite number of deep space objects, it is crucial to differentiate those space objects such as stars, galaxies, or quasars in recent or upcoming deep astronomical surveys. This task becomes very difficult because of the tedious procedure for separation between edgy and expanded sources, which makes this classification problem a difficult task. After Machine Learning approaches, there has been a rise in the use of Deep Learning methodology for deep space object classification challenges because of its improved calibration. Deep learning models such as VGG16, ResNet50, and InceptionV3 were trained using images obtained from the Slone Deep Space Survey. In addition to these pre-built architectural models, we have also implemented a CNN classifier equipped with an adam optimization parameter to explore the behavior of CNN layers. Although this CNN classifier along with other pre-trained models employed in this study was able to extract new features and classify astronomical objects with an accuracy of 79% plus, our VGG16 model achieved a significantly good accuracy of 86.04%.

Item Type: Thesis (Masters)
Uncontrolled Keywords: Deep Learning; VGG16; Pre-trained models; CNN; astronomical objects
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science
Q Science > QB Astronomy
Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: Tamara Malone
Date Deposited: 24 Jan 2023 11:09
Last Modified: 03 Mar 2023 16:48

Actions (login required)

View Item View Item