Multiple Style Transfer Via Variational Autoencoder

Modern works on style transfer focus on transferring style from a single image. Recently, some approaches study multiple style transfer; these, however, are either too slow or fail to mix multiple styles. We propose ST-VAE, a Variational AutoEncoder for latent space-based style transfer. It performs multiple style transfer by projecting nonlinear styles to a linear latent space, enabling to merge styles via linear interpolation before transferring the new style to the content image. To evaluate ST-VAE, we experiment on COCO for single and multiple style transfer. We also present a case study revealing that ST-VAE outperforms other methods while being faster, flexible, and setting a new path for multiple style transfer.

1. Introduction

Style transfer is a well-visited topic [1], [2], [3], [4]. Given a target and a reference image, its goal is to synthesize an image with content from the target and style from the reference. Typically, ‘style’ refers to color, texture, or brushstroke [5], [6].