Astronomical Object Shape Detection Using Deep Learning Models

In astronomy, over a large kilometer of distance, it has been find out that it has billions of changeable and unchangeable sources are available. The challenges is, researches are not able to split that source into necessary and unnecessary and from that source most of the shapes will never been seen through human eye. In order to overcome that the machine-learning model is designed whereas, most of the machine learning models will not provide good outcomes. In addition, it struggles to differentiate between necessary anomalies and unwanted sources like artefacts or rare anomalies source, for which the researches will not show much interest. Which say that the ML models cannot be implemented in real time. So to achieve the required outcome the deep learning models have been introduced. Which combine both the flexibility and goal of the human brain with the structure of machine learning. Space scientist do analysis on astronomy images, light curves and spectra. In our system, the transfer learning models from the deep learning have been implemented. VGGNet models like VGG16 and VGG19 models are done and then performance of both the models are compared.