MegaStyle：透過一致的文字到影像風格映射建構多樣化且可擴展的風格資料集

MegaStyle: Constructing Diverse and Scalable Style Dataset via Consistent Text-to-Image Style Mapping

April 9, 2026

作者: Junyao Gao, Sibo Liu, Jiaxing Li, Yanan Sun, Yuanpeng Tu, Fei Shen, Weidong Zhang, Cairong Zhao, Jun Zhang

cs.AI

摘要

本文提出MegaStyle——一種新穎且可擴展的數據構建流程，能夠創建具有風格內部一致性、風格間多樣性及高品質的風格數據集。我們通過利用現有大型生成模型具備的文本-圖像風格映射一致性能力實現此目標，該模型可根據給定的風格描述生成相同風格的圖像。基於此基礎，我們構建了包含17萬個風格提示詞與40萬個內容提示詞的多元化平衡提示庫，並通過內容-風格提示組合生成大規模風格數據集MegaStyle-1.4M。依托該數據集，我們提出風格監督對比學習方法來微調風格編碼器MegaStyle-Encoder，以提取具表現力的風格特徵表示，同時訓練了基於FLUX的風格遷移模型MegaStyle-FLUX。大量實驗驗證了保持風格內部一致性、風格間多樣性及高品質對於風格數據集的重要性，並證明MegaStyle-1.4M的有效性。此外，基於MegaStyle-1.4M訓練的MegaStyle-Encoder與MegaStyle-FLUX能提供可靠的風格相似度度量與泛化性強的風格遷移效果，為風格遷移領域作出重要貢獻。更多結果請參見項目網站：https://jeoyal.github.io/MegaStyle/。

English

In this paper, we introduce MegaStyle, a novel and scalable data curation pipeline that constructs an intra-style consistent, inter-style diverse and high-quality style dataset. We achieve this by leveraging the consistent text-to-image style mapping capability of current large generative models, which can generate images in the same style from a given style description. Building on this foundation, we curate a diverse and balanced prompt gallery with 170K style prompts and 400K content prompts, and generate a large-scale style dataset MegaStyle-1.4M via content-style prompt combinations. With MegaStyle-1.4M, we propose style-supervised contrastive learning to fine-tune a style encoder MegaStyle-Encoder for extracting expressive, style-specific representations, and we also train a FLUX-based style transfer model MegaStyle-FLUX. Extensive experiments demonstrate the importance of maintaining intra-style consistency, inter-style diversity and high-quality for style dataset, as well as the effectiveness of the proposed MegaStyle-1.4M. Moreover, when trained on MegaStyle-1.4M, MegaStyle-Encoder and MegaStyle-FLUX provide reliable style similarity measurement and generalizable style transfer, making a significant contribution to the style transfer community. More results are available at our project website https://jeoyal.github.io/MegaStyle/.