FRI-FRII-Simulation-using-DCGAN

Implementaion of Deep Convolutional Generative Adversarial Network using DCGAN in Tensorflow code. Detailed information can be found in this paper.

FRI and FRII sources are rare candidates. To be able to apply convolutional neural network for classification of these two sources, one needs to gather a large amount of data. Using real data of FRI and FRII sources, DCGAN has been used as a simulator. Here are examples of generated FRI and FRII sources after training the DCGAN network. This work has been submitted as part of my Master thesis: Source classification in deep radio surveys using machine learning techniques.

A Summary of my MSc Project is as follows: