Domain Adaptation using CycleGAN through style transfer

This project demonstrates a domain adaptation approach where scarce labeled data in the target domain is supplemented using labeled samples from a related source domain via style-transfer techniques (e.g., zebra-to-horse).