CSGO: Content-Style Composition in Text-to-Image Generation

TL;DR

This paper introduces IMAGStyle, a large-scale dataset for style transfer, and CSGO, a new model that improves style control in text-to-image generation through explicit content-style decoupling.

Contribution

The study presents a novel data construction pipeline, a large-scale dataset IMAGStyle, and a new style transfer model CSGO with explicit content-style separation and versatile style transfer capabilities.

Findings

IMAGStyle contains 210k stylized image triplets.

CSGO achieves improved style control in image generation.

Extensive experiments validate the effectiveness of the approach.

Abstract

The diffusion model has shown exceptional capabilities in controlled image generation, which has further fueled interest in image style transfer. Existing works mainly focus on training free-based methods (e.g., image inversion) due to the scarcity of specific data. In this study, we present a data construction pipeline for content-style-stylized image triplets that generates and automatically cleanses stylized data triplets. Based on this pipeline, we construct a dataset IMAGStyle, the first large-scale style transfer dataset containing 210k image triplets, available for the community to explore and research. Equipped with IMAGStyle, we propose CSGO, a style transfer model based on end-to-end training, which explicitly decouples content and style features employing independent feature injection. The unified CSGO implements image-driven style transfer, text-driven stylized synthesis, and text editing-driven stylized synthesis. Extensive experiments demonstrate the effectiveness of our approach in enhancing style control capabilities in image generation. Additional visualization and access to the source code can be located on the project page: \url{https://csgo-gen.github.io/}.