[论文] Tstars-Tryon 1.0: Robust and Realistic Virtual Try-On for Diverse Fash...

## 论文概要 **研究领域**: CV **作者**: Mengting Chen, Zhengrui Chen, Yongchao Du, Zuan Gao, Taihang Hu, Jinsong Lan, Chao Lin, Yefeng Shen, Xingjian Wang, Zhao Wang, Zhengtao Wu, Xiaoli Xu, Zhengze Xu, Hao Yan, Mingzhou Zhang, Jun Zheng, Qinye Zhou, Xiaoyong Zhu, Bo Zheng **发布时间**: 2026-04-21 **arXiv**: [2604.19748](https://arxiv.org/abs/2604.19748) ## 中文摘要 近年来，图像生成与编辑技术的进步为虚拟试衣开辟了新机遇。然而，现有方法仍难以满足复杂的现实需求。我们推出了 Tstars-Tryon 1.0——一个具备商业级规模的虚拟试衣系统，兼具鲁棒性、真实感、通用性与高效性。首先，系统在极端姿态、剧烈光照变化、运动模糊等野外挑战性场景下保持高成功率；其次，生成结果具有高度照片级真实感与细粒度细节，忠实保留服装纹理、材质属性与结构特征，同时有效避免常见 AI 伪影；第三，除服装试穿外，模型支持多达 6 张参考图像的灵活多图合成，覆盖 8 大时尚品类，并协同控制人物身份与背景；第四，针对商业部署的延迟瓶颈，系统经深度推理优化，实现近实时生成，确保流畅用户体验。上述能力源于端到端模型架构、可扩展数据引擎、鲁棒基础设施与多阶段训练范式的集成设计。大量评估与大规模产品部署表明，Tstars-Tryon 1.0 达到业界领先的整体性能。为支持未来研究，我们还发布了全面基准测试。该模型已在淘宝 App 实现工业级部署，服务数百万用户，日处理数千万级请求。 ## 原文摘要 Recent advances in image generation and editing have opened new opportunities for virtual try-on. However, existing methods still struggle to meet complex real-world demands. We present Tstars-Tryon 1.0, a commercial-scale virtual try-on system that is robust, realistic, versatile, and highly efficient. First, our system maintains a high success rate across challenging cases like extreme poses, severe illumination variations, motion blur, and other in-the-wild conditions. Second, it delivers highly photorealistic results with fine-grained details, faithfully preserving garment texture, material properties, and structural characteristics, while largely avoiding common AI-generated artifacts. Third, beyond apparel try-on, our model supports flexible multi-image composition (up to 6 reference... --- *自动采集于 2026-04-23* #论文 #arXiv #CV #小凯