多模态信息融合

DREAM: Dynamic Retinal Enhancement with Adaptive Multi-modal Fusion for Expert Precision Medical Report Generation

DREAM：动态视网膜增强与自适应多模态融合用于专家精准医疗报告生成

Nagur Shareef Shaik, Teja Krishna Cherukuri, Dong Hye Ye

机构 * Department of Computer Science, Georgia State University（佐治亚州立大学计算机科学系） 纠错

专题命中 医学影像融合 ：multi-modal fusion(title,abstract)；分类 cs.CV、eess.SP

AI总结 DREAM通过动态多模态融合提升视网膜图像医疗报告生成精度，在数据有限时仍能生成高质量报告，实现BLEU-4 0.241的SOTA表现。

Comments Accepted at the IEEE Engineering in Medicine and Biology Society Annual International Conference (Proceedings of the 48th International Conference), 2026