InsertFuse：多类别参考引导图像插入的统一框架
原标题：InsertFuse: A Unified Framework for Multi-Category Reference-Guided Image Insertion
AI 摘要
arXiv 论文提出 InsertFuse，一个用于多类别参考引导图像插入的统一框架。其核心思想是将类别特定专家学习与跨类别能力整合解耦，通过插入在策略蒸馏（IOPD）将多个专家能力整合到单一学生模型中。此外，引入 Token 对齐几何条件（TAGC）和区域平衡流匹配以提升空间控制，以及参考 CFG 增强参考引导。在 AnyInsertion 基准和自建多类别测试集上取得了最先进性能。
正文节选
Computer Science > Computer Vision and Pattern Recognition Title:InsertFuse: A Unified Framework for Multi-Category Reference-Guided Image Insertion View PDF HTML (experimental) Abstract:We present InsertFuse, a unified framework for multi-category reference-guided image insertion. Its key idea is to decouple category-specific expertise learning from cross-category capability consolidation. InsertFuse first trains specialized experts for different insertion categories and then introd