Translate the Facial Regions You Like Using Region-Wise Normalization

TL;DR

This paper introduces a region-wise normalization framework for face translation that enables precise, region-specific style transfer, improving results over existing methods and allowing detailed control over facial region modifications.

Contribution

The paper proposes a novel region-wise normalization block and a region matching loss for improved, controllable face translation at the region level using GANs.

Findings

Significant improvement over state-of-the-art methods like StarGAN, SEAN, and FUNIT.

Enhanced control over specific facial regions during translation.

Effective translation of shape and texture for individual facial regions.

Abstract

Though GAN (Generative Adversarial Networks) based technique has greatly advanced the performance of image synthesis and face translation, only few works available in literature provide region based style encoding and translation. We propose in this paper a region-wise normalization framework, for region level face translation. While per-region style is encoded using available approach, we build a so called RIN (region-wise normalization) block to individually inject the styles into per-region feature maps and then fuse them for following convolution and upsampling. Both shape and texture of different regions can thus be translated to various target styles. A region matching loss has also been proposed to significantly reduce the inference between regions during the translation process. Extensive experiments on three publicly available datasets, i.e. Morph, RaFD and CelebAMask-HQ, suggest that our approach demonstrate a large improvement over state-of-the-art methods like StarGAN, SEAN and FUNIT. Our approach has further advantages in precise control of the regions to be translated. As a result, region level expression changes and step by step make up can be achieved. The video demo is available at https://youtu.be/ceRqsbzXAfk.