題 目：Variational Tensor-based Models for Image Diffusion
報告人：Freddie Astrom （Linkoping University, Sweden）
時 間：2014年1月26日（周一）9:00
地 點：燕山校區1號教學樓1406室
報告主要內容：
Inverse problems often arise in the context of information reconstruction from a set of known, measurable, and possibly noisy data. Even though inverse problems have been extensively studied, many theoretical and practical problems remain unsolved. In this talk we will briefly review classical mathematical models for inverse problems and present methods for information recovery in image processing applications. Particularly, we consider the problem of image denoising and the importance of regularization. We concentrate on Tikhonov-type regularization, which is subject to an extensive discussion from an image analysis perspective. While giving an overview of established diffusion formulations we, formulate and present, tensor-based variational models with a focus on image enhancement.
