Projects

Trio Diffusion

My Experiment with Infinite Image Generation

TailorGAN

Generating Infinite-Length Images with a Relatively Small-sized Conditional Generative Adversarial Network Variant

Unlocking Facial Recognition Systems with Master-Key Images

Generalized Attacks on Face Verification Systems

AutoGAN

An Automated Human-out-of-the-Loop Approach for Training Generative Adversarial Networks

Exploring GAN-Based Oversampling across Varied Data Difficulty Factors

This research examines the effects of data difficulty factors on GAN-based upsampling in imagery and cybersecurity data. Factors include imbalance ratio, sample size, data dimensionality, and class overlap. The study aims to understand their impact on performance.