Click:

The Last Update Time:--

Guobin Zhu
+
Personal Information
  • Supervisor of Doctorate Candidates
  • Supervisor of Master's Candidates
  • Name (Pinyin):Zhu Guobin
  • E-Mail:
  • Education Level:研究生毕业
  • Business Address:Room 1104, School of Cyber Science and Engineering, Wuhan University
  • Gender:Male
  • Contact Information:Tel. +86-27-68778916 Mobile: +86-18040500277 http://rsgis.whu.edu.cn/index.php?m=content&c=index&a=show&catid=127&id=6311
  • Academic Titles:Director of Dept. of Spatial Information & Digital Technology
  • Alma Mater:Ben-Gurion University of the Negev
  • Teacher College:School of Remote Sensing and Information Engineering
  • Discipline: Photogrammetry and Remote Sensing
  • Honors and Titles:
      2018  elected:Honorary Title of Hubei Provincial Industrial Professors

      2017  elected:Honorary Title of Innovation & Creative Talent of Yangzhou City, Jiangsu Province

      2016  elected:Honorary Title of Outstanding Communist Party Member of Wuhan University

      2016  elected:Honorary Title of Progressive Individual for Major Evaluation at Wuhan University

      2015  elected:湖北省优秀学士论文指导教师

      2015  elected:Honorary Title of Outstanding Navigator for Students of Wuhan University

      2014  elected:Honorary Title of the Yellow Crane Talent of Wuhan City

      2013  elected:Honorary Title of Representatives of the CPC Congress of Wuhan University

      2013  elected:Honorary Title of Progressive Individual for Teaching and Educating at Wuhan University
Other Contact Information
  • ZipCode:
  • PostalAddress:
  • OfficePhone:
  • Telephone:
  • email:
Current position: Home   >   Student Information

付育荣

  • Date:2018-12-03
  • Hits:
  • Duration of Study: 

    2016-2018
  • Major: 

    Software Engineering
  • Research Focus: 

    通信与智能工程
  • Personal Profile: 


    将深度学习理论应用于图像超分辨率问题的学习与研究,提出了一种基于高低频针对性深度学习的人脸超分辨率重构算法。该算法通过利用深度卷积神经网络将低分辨率图像分为高频和低频两个区域并分别对它们进行特征学习和增强,最后重构出更为精确的与其对应的高分辨率图像。具体来说,该网络分为高低频区域提取模块、高低频特征增强模块和高低频特征融合模块三个部分。首先,该网络通过高低频区域提取模块将图片中对应的高低频区域划分出来;随后在高低频特征增强模块中对提取出的区域做针对性增强;最后,在高低频特征融合模块中对提取出的人脸特征进行融合,重建出最终的高分辨率人脸。
    本文的算法将人脸图像的高频细节和低频细节分开增强,使得重建图片针对不同频率区域都有更好的效果。与已有算法相比,本文算法克服了已有算法对高频区域细节重建效果模糊的问题,在图像超分辨率算法通用的评价数据集上的实验效果优于之前的超分辨率算法,重构出更为清晰的轮廓区域。
  • Education Level: 

    研究生毕业
  • Degree: 

    硕士
  • Current Status: 

    离校
  • Graduation Thesis Title: 

    《基于深度学习的人脸超分辨率重构算法研究》
  • Student ID: 

    2016282160037
  • Date of Registration: 

    2016-09-01
  • Date of Graduation: 

    2018-06-30