• Version
• Download 3
• File Size 366.92 KB
• File Count 1
• Create Date 5 June 2025
• Last Updated 5 June 2025

A Deep Learning Approach for Detecting Unusual Celestial Phenomena in Astronomical Datasets

Authors:

Mr. Mula Mahender^*1, Samala Abhinav^*2, Sulthani Varun^*3, Garishe Mahender^*4

^*1Associate Professor of Department of CSE (AI & ML) Of ACE Engineering College, India.

^*2,3,4 Department CSE (AI & ML) Of ACE Engineering College, India.

Abstract: Detection of Unusual Celestial Phenomena in Astronomical Datasets is a deep learning -powered project implemented using ResNet50, a Convolutional Neural Network (CNN) model, and TensorFlow. Utilizing these it efficiently scans big open-source datasets such as NASA (National Aeronautics and Space Administration), SDSS (Sloan Digital Sky Survey), ESA (European Space Agency), etc, to identify asteroids, galaxies, exoplanets, supernovae, and other irregularities. Conventional human analysis is time-consuming and prone to errors, whereas current automated solutions are not affordable and scalable. Our solution combines FastAPI and cloud hosting (Render) for real-time use, making it affordable for students, researchers, and startups. The system uses consumer GPUs, making space research more democratic with an affordable, scalable, and open-source approach.

Keywords: Deep learning, ResNet50, TensorFlow, FastAPI, cloud deployment.