报告题目:Geometric Priors: The bridge between natural images and biomedical images
主讲嘉宾:龚元浩 助理教授  深圳大学
邀请人:邱国平教授
时间:2018年 11月 22日(周 四)下午4:00
地点:深圳大学南校区基础实验楼北座信息工程学院N710会议室
报告摘要:
Priors play fundamental roles in various image processing and deep learning tasks. Geometric priors are particularly interesting because of their intuitive explanation. Minimizing these priors, however, is challenging. In this talk, I will show to model such geometric priors and their related optimization algorithms. More specifically, I will show how to use simple filters to minimize curvatures. These filters are 1000 or 10,000 faster than traditional methods.

嘉宾简介:
 龚元浩,深圳大学助理教授,2007年本科毕业于清华大学,2015年博士毕业于瑞士苏黎世联邦理工学院,主要从事图像处理、医学图像分析和计算机视觉方面的研究。