医学 AI

Decentralized LoRA augmented transformer with multi-scale feature learning for secured eye diagnosis

去中心化LoRA增强变换器与多尺度特征学习用于安全眼疾诊断

Md. Naimur Asif Borno, Md Sakib Hossain Shovon, MD Hanif Sikder, Iffat Firozy Rimi, Tahani Jaser Alahmadi, Mohammad Ali Moni

机构 * organization= Research Assistant, The University of Queensland , addressline= 308 Queen St , city= Brisbane City , postcode= QLD 4000 , state= Queensland , country= Australia ； organization= Mechatronics Engineering, Rajshahi University of Engineering \& Technology , city= Rajshahi , postcode= 6204 , country= Bangladesh ； organization= Researcher, The University of Queensland , addressline= 308 Queen St , city= Brisbane City , postcode= QLD 4000 , state= Queensland , country= Australia ； organization= Department of Computer Science, American International University Bangladesh , city= Dhaka , postcode= 1216 , country= Bangladesh ； organization= Department of Computer Science, University of South Asia-Bangladesh , city= Dhaka , postcode= 1216 , country= Bangladesh ； organization= Department of Computer Science ； Engineering, Daffodil International University , city= Dhaka , country= Bangladesh ； Department of Information Systems, College of Computer ； Information Sciences, Princess Nourah bint Abdulrahman University, P.O. Box 84428, Riyadh, Saudi Arabia. Email ； organization= Faculty of Health, Medicine ； Behavioural Sciences, The University of Queensland , addressline= 308 Queen St , city= Brisbane City , postcode= QLD 4000 , state= Queensland , country= Australia ； Cyber Futures Institute Charles Sturt University , addressline= 308 Queen St , city= Bathurst NSW , country= Australia 纠错

专题命中 医学数据与评测 ：diagnosis(title,abstract)；分类 cs.CV

AI总结 本文提出一种高效图像变换器框架，结合多尺度特征学习、LoRA、知识蒸馏和联邦学习，解决医学影像中疾病诊断的准确性和隐私保护问题，实验表明其在AUC、F1分数和精度上优于传统CNN和现有Transformer架构。

Comments Published at Knowledge-Based Systems

GenShield: Unified Detection and Artifact Correction for AI-Generated Images

GenShield：面向AI生成图像的统一检测与伪影校正

Zhipei Xu, Xuanyu Zhang, Youmin Xu, Qing Huang, Shen Chen, Taiping Yao, Shouhong Ding, Jian Zhang

机构 * School of Electronic and Computer Engineering, Peking University（北京大学电子与计算机工程学院） ； Tencent Youtu Lab（腾讯优图实验室） 纠错

专题命中 医学数据与评测 ：diagnosis(abstract)；分类 cs.CV

AI总结 本文提出GenShield框架，通过闭环诊断与修复流程实现可解释的AI生成图像检测与可控伪影校正，结合视觉链式推理课程学习策略，提升校正效果与泛化能力。