High-Performance Dual Optimization Methods to Medical Image Analysis
发布日期：2015-07-14 11:26:06 浏览次数：次
Robarts Research Institute, Western University
Abstract: Many problems of medical image analysis are challenging due to the associated complex optimization formulations and constraints, extremely big image data volumes to be processed, poor imaging quality, missing data etc. On the other hand, it is highly desired in clinical practices to process and analyze the acquired imaging data, for example segmentation and registration, in an automated and efficient numerical way, which motivated vast active studies during the last 30 years, in a rather broad sense, to deliver advanced mathematical analysis and develop high-performance numerical schemes. This talk targets to present an overview of dual optimization theories, which becomes one most successful optimization framework of image processing developed recently, with a wide spectrum of applications to medical image analysis. It focuses on the optimization problems arising from two most interesting topics: medical image segmentation and registration, and presents both analysis and algorithmic solutions in a unified manner in terms of dual optimization.