澳门新葡8455最新网站,www.8455.com,新葡京最新官网

当前位置: 澳门新葡8455最新网站 > 学术活动 > 正文
Regularized Variational Model with Deep Learning for Image Dehazing
澳门新葡8455最新网站:2019年12月09日 09:27 点击数:

报告人:刘文

报告地点:澳门新葡8455最新网站104报告厅

报告澳门新葡8455最新网站:2019年 12月12日星期四15:00-16:00

邀请人:刘俊

报告摘要:

Maritime images captured under hazy weather conditions commonly suffer from reduced contrast and visibility. High-quality dehazing performance is highly dependent upon the accurate estimation of transmission map. In this talk, we introduced two regularized variational models for image dehazing. (1) In the first model, the coarse transmission map is first obtained by weightedly fusing two different transmission maps, which are generated from foreground and sky regions, respectively. A hybrid variational model with promoted regularization terms is then proposed to assisting in refining transmission map. The final haze-free image can be effectively obtained according to the refined transmission map and atmospheric scattering model. (2) In the second model, we first rewrite and simplify the original imaging degradation model to assist in estimating transmission map. A regularized variational model with deep learning is then developed to restore the degraded images. The proposed framework has the capacity of improving image quality while suppressing the unwated artifacts in practical applications.

主讲人概况:

刘文,武汉理工大学副教授,博士生导师,湖北省“楚天学者计划”入选者,深圳大学访问学者(2016.04-2017.10),新加坡科技研究局访问学者(2018.03)。2009年本科毕业于武汉理工大学信息与计算科学专业,2015年获香港中文大学哲学博士学位,中国科新葡京最新官网自动化研究所访问学生。主要研究方向为视觉感知与智能计算、海事大数据挖掘与可视分析。近期主要承担国家自然科学基金青年项目1项,国家重点研发计划子课题和军队后勤开放研究项目各1项。近5年在SCI期刊与EI会议上发表论文50余篇;研究成果获2018年中国水运建设行业协会科学技术奖一等奖,2019年机器学习和计算国际会议(ICMLC)最佳报告奖与中国航海科技期刊论文二等奖;目前担任SCI期刊International Journal on Semantic Web and Information Systems (CCF-C)副主编和SCI期刊Sensors客座编辑(Guest Editor);以色列科技部基金项目和国家自然科学基金项目评议专家。

Copyright ©版权所有:澳门新葡8455最新网站

地址:吉林省长春市人民大街5268号|邮编:130024|电话:0431-85099589|传真:0431-85098237


XML 地图 | Sitemap 地图