XGAN: UNSUPERVISED IMAGE-TO-IMAGE TRANSLATION FOR MANY-TO-MANY MAPPINGS

https://arxiv.org/pdf/1711.05139.pdf

Novelty

• Unsupervise
• Two domain share the same latent space
• predict class at encoder

Problem

• Only two domains at a time, need to train new model for new domain

Structure

Structure.png

Loss

Loss.png

Dataset

Self-created CartoonSet dataset