Augmented CycleGAN: Learning Many-to-Many Mappings from Unpaired Data

TL;DR

Augmented CycleGAN introduces a model capable of learning many-to-many domain mappings from unpaired data, enhancing flexibility for structured prediction tasks like image segmentation.

Contribution

It extends CycleGAN to handle many-to-many mappings, overcoming the deterministic, one-to-one assumption of previous models.

Findings

Successfully learns complex many-to-many mappings

Outperforms previous models on multiple datasets

Demonstrates improved structured prediction results

Abstract

Learning inter-domain mappings from unpaired data can improve performance in structured prediction tasks, such as image segmentation, by reducing the need for paired data. CycleGAN was recently proposed for this problem, but critically assumes the underlying inter-domain mapping is approximately deterministic and one-to-one. This assumption renders the model ineffective for tasks requiring flexible, many-to-many mappings. We propose a new model, called Augmented CycleGAN, which learns many-to-many mappings between domains. We examine Augmented CycleGAN qualitatively and quantitatively on several image datasets.