Few-Shot Unsupervised Image-to-Image Translation

TL;DR

This paper introduces a few-shot unsupervised image-to-image translation method that can generalize to unseen classes with only a few example images, using a novel network design and adversarial training.

Contribution

It proposes a new framework for few-shot unsupervised image translation capable of handling unseen classes with minimal data, advancing beyond existing methods requiring many images.

Findings

Effective few-shot translation on benchmark datasets

Outperforms baseline methods in quality and generalization

Demonstrates strong generalization to unseen classes

Abstract

Unsupervised image-to-image translation methods learn to map images in a given class to an analogous image in a different class, drawing on unstructured (non-registered) datasets of images. While remarkably successful, current methods require access to many images in both source and destination classes at training time. We argue this greatly limits their use. Drawing inspiration from the human capability of picking up the essence of a novel object from a small number of examples and generalizing from there, we seek a few-shot, unsupervised image-to-image translation algorithm that works on previously unseen target classes that are specified, at test time, only by a few example images. Our model achieves this few-shot generation capability by coupling an adversarial training scheme with a novel network design. Through extensive experimental validation and comparisons to several baseline methods on benchmark datasets, we verify the effectiveness of the proposed framework. Our implementation and datasets are available at https://github.com/NVlabs/FUNIT .