Region-Aware Multimodal Large Language Model via SlowFast Tokenization and Pseudo-Mask Guidance for 3D CT Report Generation
区域感知多模态大语言模型:基于慢快标记化与伪掩码引导的3D CT报告生成
机构 * Department of Convergence Medicine, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Republic of Korea(韩国首尔峨山医疗中心蔚山大学医学院融合医学系) ; University of Ulsan College of Medicine, Seoul, Republic of Korea(韩国首尔蔚山大学医学院) ; Department of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Republic of Korea(韩国首尔峨山医疗中心蔚山大学医学院放射科与放射学研究所)
专题命中 领域大模型 :large language model(title,abstract);language model(title,abstract)
AI总结 提出MedRegion-CT框架,通过区域慢快标记器联合建模全局与细粒度信息、伪掩码引导关注诊断关键区域、结构化病变信息提示,实现高质量CT报告生成,在多项指标上达到最优。
Comments Accepted to ECCV 2026. 15 pages, 8 figures, 4 tables