TransferI2I: Transfer Learning for Image-to-Image Translation from Small Datasets

TL;DR

TransferI2I introduces a novel transfer learning approach for image-to-image translation that effectively handles small datasets by decoupling generation and translation steps, utilizing innovative initialization techniques and an auxiliary GAN.

Contribution

The paper proposes a new transfer learning framework for I2I translation that improves performance on small datasets through source-target and self-initialization techniques, and an auxiliary GAN.

Findings

Outperforms existing methods on three datasets.

Achieves over 25 points improvement in mFID.

Effective training from small datasets.

Abstract

Image-to-image (I2I) translation has matured in recent years and is able to generate high-quality realistic images. However, despite current success, it still faces important challenges when applied to small domains. Existing methods use transfer learning for I2I translation, but they still require the learning of millions of parameters from scratch. This drawback severely limits its application on small domains. In this paper, we propose a new transfer learning for I2I translation (TransferI2I). We decouple our learning process into the image generation step and the I2I translation step. In the first step we propose two novel techniques: source-target initialization and self-initialization of the adaptor layer. The former finetunes the pretrained generative model (e.g., StyleGAN) on source and target data. The latter allows to initialize all non-pretrained network parameters without the need of any data. These techniques provide a better initialization for the I2I translation step. In addition, we introduce an auxiliary GAN that further facilitates the training of deep I2I systems even from small datasets. In extensive experiments on three datasets, (Animal faces, Birds, and Foods), we show that we outperform existing methods and that mFID improves on several datasets with over 25 points.