arXivDaily arXiv每日学术速递 周一至周五更新

高校专区

Beihang University(北京航空航天大学)

2025-12-17 至 2025-12-17 共收录 3
2512.14550 2025-12-17 cs.CV

TAT: Task-Adaptive Transformer for All-in-One Medical Image Restoration

TAT:适用于所有医学图像恢复的任务自适应Transformer

Zhiwen Yang, Jiaju Zhang, Yang Yi, Jian Liang, Bingzheng Wei, Yan Xu

机构 * School of Biological Science and Medical Engineering(生物科学与医学工程学院) State Key Laboratory of Software Development Environment(软件开发环境国家重点实验室) Key Laboratory of Biomechanics and Mechanobiology of Ministry of Education(教育部生物力学与机力学重点实验室) Beijing Advanced Innovation Center for Biomedical Engineering(北京生物医学工程先进创新中心) Beihang University(北京航空航天大学) Sinovision Technologies (Beijing) Co., Ltd.(北京视界科技有限公司) ByteDance Inc.(字节跳动公司)

AI总结 TAT提出了一种任务自适应Transformer框架,通过动态调整权重和损失平衡策略,解决医学图像恢复中任务干扰和不平衡问题,实现多任务高性能恢复。

Comments This paper has been accepted by MICCAI 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2511.01510 2025-12-17 cs.CV

Luminance-Aware Statistical Quantization: Unsupervised Hierarchical Learning for Illumination Enhancement

具有亮度意识的统计量化:无监督分层学习用于照明增强

Derong Kong, Zhixiong Yang, Shengxi Li, Shuaifeng Zhi, Li Liu, Zhen Liu, Jingyuan Xia

机构 * College of Electronic Science and Technology, National University of Defense Technology(电子科学与技术学院,国防科技大学) College of Electronic and Information Engineering, Beihang University(电子与信息工程学院,北航)

AI总结 本文提出LASQ框架,通过无监督分层学习实现照明增强,改进亮度转换建模,提升低光照图像恢复的适应性和泛化能力。

Comments Accepted at NeurIPS 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.14158 2025-12-17 cs.CV cs.CR

CIS-BA: Continuous Interaction Space Based Backdoor Attack for Object Detection in the Real-World

CIS-BA:基于连续交互空间的面向现实世界目标检测的后门攻击

Shuxin Zhao, Bo Lang, Nan Xiao, Yilang Zhang

机构 * organization= State Key Laboratory of Complex \& Critical Software Environment , addressline= Beihang University , city= Beijing , postcode= 100191 , country= China organization= Zhongguancun Laboratory , addressline= , city= Beijing , postcode= 100080 , country= China

AI总结 CIS-BA通过引入连续交互空间的触发机制,实现了对现实世界目标检测模型的高效且鲁棒的后门攻击,有效提升了攻击成功率并规避了现有防御方法。

详情

展开后加载摘要…

URL PDF HTML 收藏