Using a GAN to synthetically generate medical images for DL purposes

A data-driven framework was developed to predict and optimize the pressure-surface film cooling effectiveness of the GE-EEE first-stage turbine using deep learning.

Conditional Generative Adversarial Networks (CGANs) extend the capabilities of traditional GANs by conditioning both the generator and discriminator models on additional information, typically class labels or other forms of auxiliary information.

Bachelor Degree Project in Information Technology

Utilized a VAE (Variational Autoencoder) and CGAN (Conditional Generative Adversarial Network) models to generate synthetic chatter signals, addressing the challenge of imbalanced data in turning operations. Compared othe performance of synthetic chatter signals.

Training cGAN's using a small set of labelled samples + large number of unlabelled samples