通过分阶段持续学习，实现风格迁移中内容与风格的精准分离。
Style-CCL: Content-Preserving Style Transfer via Curriculum Continual Learning
- 分阶段从语义到纹理风格逐步训练，避免风格主导内容
- 在百万级数据上实现风格相似度、内容一致性双提升
- 适合需要高保真内容还原的图像生成应用
给定内容和风格参考图，基于扩散变换器（DiTs）的内容保持风格迁移仍面临挑战，因内容与风格特征纠缠。我们构建了百万规模的反向三元组合成数据集，并设计双分支风格-内容扩散变换器（SC-DiT），通过独立的ROPE嵌入与因果掩码解耦风格与内容。然而，单一阶段混合风格类别训练导致语义风格主导，抑制纹理风格学习并损害内容保留。为此，提出多阶段课程持续学习框架Style-CCL，按从语义（易）到纹理（难）风格、从干净到合成数据的顺序训练，跨阶段采用随机记忆回放以避免灾难性遗忘。大量实验表明，Style-CCL在三个核心指标——风格相似度、内容一致性和美学质量——上达到当前最优性能。
原文摘要 · Abstract (English)
Content-Preserving Style transfer, given content and style references, remains challenging for Diffusion Transformers (DiTs) due to entangled content and style features. With a reverse triplet synthesis pipeline to build a million-scale training set and a dual-branch Style-Content DiT (SC-DiT) that decouples style and content via separate ROPE embeddings and causal masking, we observe that such a one-stage training paradigm on mixed style categories causes semantic styles to dominate, hindering texture style learning, and harming content preservation. To address these issues, we propose Style-CCL, a Multi-Stage Curriculum Continual Learning framework that trains SC-DiT from semantic (easy) to texture (hard) styles, and from clean to synthetic data, with Random Memory Rehearsal across stages to avoid catastrophic forgetting. Extensive experiments demonstrate that our Style-CCL achieves state-of-the-art performance in three core metrics: style similarity, content consistency, and aesthetic quality.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。