报告人:袁景
工作单位:西安电子科技大学
报告时间:12月27日9:30
报告地点:数学院北研教室
报告摘要:
Many problems of medical image analysis are challenging due to the associated complex optimization formulations and constraints, extremely big image data being processed, poor imaging quality, missing data etc. On the other hand, it is highly desired to process and analyze the acquired imaging data, for example segmentation and registration etc., in an automated and efficient numerical way, which motivated vast active studies during the last 30 years, in a rather broad sense. This talk targets to present an overview of modern dual optimization theory, which delivers an advanced unified framework of mathematical analysis and high-performance ADMM numerical schemes along with a wide spectrum of applications. We focus on the optimization problems arising from the most interesting topics:segmentation and registration, and present both analysis and high-performance numerical solutions in a unified manner in terms of dual optimization.
报告人简介:
西安电子科技大学特聘教授。本科和博士分别毕业于北京大学物理学院和德国海德堡大学数学与计算机科学学院,后在加拿大西安大略大学从事博士后研究,加拿大Robarts 研究所研究员,现为魁北克省大学和南方科技大学客座教授,从事最优化理论与算法研究、并行化快速优化算法实现及其在计算机医学图像分析、计算视觉与机器学习的相关应用,指导硕士、博士研究生以及博士后的研究工作,在IJCV,CVPR,ECCV, ICML, IEEE TIP, IEEE PAMI等相关顶级学术刊物发表论文近100篇。