Portrait Stylization: Artistic Style Transfer with Auxiliary Networks for Human Face Stylization

TL;DR

This paper introduces a face recognition-based auxiliary network to improve the preservation of individual facial features during artistic style transfer, addressing limitations of existing methods.

Contribution

It proposes integrating embeddings from a pre-trained face recognition model to enhance facial feature retention in style transfer.

Findings

Improved facial feature preservation in stylized images

Enhanced style transfer quality for human faces

Demonstrated effectiveness on various face images

Abstract

Today's image style transfer methods have difficulty retaining humans face individual features after the whole stylizing process. This occurs because the features like face geometry and people's expressions are not captured by the general-purpose image classifiers like the VGG-19 pre-trained models. This paper proposes the use of embeddings from an auxiliary pre-trained face recognition model to encourage the algorithm to propagate human face features from the content image to the final stylized result.