讲座题目:Image retinex based on the nonconvex TV-type regularization
主讲人:庞志峰 河南大学副教授
讲座时间:12月2日(周四)下午3:00-4:00
讲座地点:
线下:综合楼601 会议室
线上:腾讯会议:https://meeting.tencent.com/dm/7Wr4vQcomfDN
会议ID:腾讯会议:265-291-589
主讲人简介:
庞志峰,博士, 河南大学副教授, 硕士生导师。河南省应用数学中心副主任, 目前兼任河南省数字图形图像学会常务理事和常务秘书长, 河南省数字图形图像学会精准放疗专委会副主委,主要研究医学图像处理和机器学习中的数学理论与数值算法。主持国家自然科学基金1项, 参与国家自然科学基金4项, 国家重点基础研究发展计划(973项目)1项。现发表相关学术论文40余篇。近年来,并受邀境内外多地访问:2010.06-2011.03, 南洋理工大学数学物理学院, 博士后; 2011.12-2012.12, 香港城市大学电脑科学系, 博士后; 2014.02月-2014.08, 中国科学院数学与系统科学研究生院, 访问学者; 2015.07-2015.08, 南开大学陈省身数学研究所, 访问学者; 2016.01-2016.02, 香港中文大学系统工程与工程管理系, 访问学者;2016.08-2017.08, 利物浦大学(英国), 数学科学学院, 访问学者; 2018.01-2018.02, 香港理工大学访问学者。
讲座摘要:
Retinex theory is introduced to show how the human visual system perceives the color and the illumination effect such as retinex illusions, medical image intensity inhomogeneity and color shadow effect etc.. Many researchers have studied this ill-posed problem based on the framework of the variation energy functional for decades. However, to the best of our knowledge, the existing models via the sparsity of the image based on the nonconvex lp-quasinorm were limited. To deal with this problem, this paper considers a TVp-HOTVq-based retinex model with $p,q/in(0; 1).$ Specially, the TVp term based on the total variation(TV) regularization can describe the reflectance efficiently, which has the piecewise constant structure. The HOTVq term based on the high order total variation(HOTV) regularization can penalize the smooth structure called the illumination. Since the proposed model is non-convex, non-smooth and non-Lipschitz, we employ the iteratively reweighed L1 (IRL1) algorithm to solve it. We also discuss some properties of our proposed model and algorithm. Experimental experiments on the simulated and real images illustrate
the effectiveness and the robustness of our proposed model both visually and quantity.
Joint with Yuan Wang, Yuping Duan and Ke Chen.