理学院“格物论坛”学术报告(十八-二十三）

理学院“格物论坛”学术报告(十八-二十三）

 

报告题目： Mathematical modeling and solution of different imaging and supervised learning problems, using variational techniques.

Topics to cover:

Introduction to variational techniques

Image denoising and inpainting

Image registration

Introduction to level set methods and image segmentation

Demodulating phase maps from fringe pattern images

Supervised learning 

Numerical methods

报 告 人：Universidad Atuonoma de Yucatan

       Carlos Francisco Brito Loeza  教授、博导

时    间：

5月3号 周四 10:00-11:00；5月4号 周五 10:00-11:00

5月5号 周六 10:00-11:00；5月6号 周日 10:00-11:00

5月7号 周一 14:30-15:30；5月8号 周二 14:30-15:30

地    点：理生楼多功能厅B705

简介：

In the last two decades, variational techniques have proved to be successful in solving difficult problems in the image restoration and processing fields. Examples of such techniques are image denoising, inpainting, segmentation and registration algorithms. More recently, the same techniques are being explored in the fringe pattern analysis field and for developing new supervised learning methods. In these series of talks we will cover the basic foundations of the variational methods, then move to some applications and finally we will explore some popular numerical methods.。

欢迎广大师生参加！

           南昌大学科技处

          理学院数学系

           二〇一八年五月