Feihu Zhang  

PhD Candidate

Department of Engineering Science
University of Oxford


Email: feihu.zhang [at] eng.ox.ac.uk

Biography

Currently, I am a first year DPhil/PhD student in Torr Vision Group and Active Vision Lab at Univeristy of Oxford, supervised by Prof. Victor Prisacariu and Prof. Philip H. S. Torr. My research interests include computer vision, deep learning and autonomous driving. I got my MPhil Degree from The Chinese Univeristy of Hong Kong under the guidance of Prof. Benjamin Wah. Then, I worked with Prof. Ruigang Yang at Baidu's Robotics and Autonomous Driving Lab (RAL) for about one year.

News

Education

Employment

Publications [My Google Scholar]

GANet: Guided Aggregation Net for End-to-end Stereo matching
Feihu Zhang, Victor Prisacariu, Ruigang Yang, Philip H.S. Torr.
IEEE Conference on Computer Vision and Pattern Recognition(CVPR), 2019.
Oral Presentation
(project  code  slides  video)
 
Hypergraph Convolution and Hypergraph Attention
Song Bai, Feihu Zhang, Philip H.S. Torr.
arXiv preprint arXiv:1901.08150 (2019).
 
Supplementary Meta-Learning: Towards a Dynamic Model for Deep Neural Networks
Feihu Zhang, Benjamin W. Wah.
IEEE International Conference on Computer Vision (ICCV), 2017.
(project  code)
 
Principles on Learning New Features for Effective Dense Matching
Feihu Zhang, Benjamin W. Wah.
IEEE Transaction on Imaging Processing (TIP), 2016.
(project  code)
 
Segment Graph Based Image Filtering: Fast Structure-Preserving Smoothing
Feihu Zhang, Longquan Dai, Shiming Xiang, Xiaopeng Zhang.
IEEE International Conference on Computer Vision (ICCV), 2015.
(project  code  video  supplementary)
 
Fully Connected Guided Image Filtering
Longquan Dai, Yuanke Meng, Feihu Zhang, Xiaopeng Zhang.
IEEE International Conference on Computer Vision (ICCV), 2015.
(project  video)
 
PM-PM: PatchMatch with Potts Model for Object Segmentation and Stereo Matching
Shibiao Xu, Feihu Zhang, Xiaofei He, Xiaopeng Zhang.
IEEE Transaction on Imaging Processing (TIP) 24.7 (2015): 2182 - 2196.
(project)
 
Fast Minimax Path-Based Joint Depth Interpolation
Longquan Dai, Feihu Zhang, Xing Mei.
IEEE Signal Processing Letters 22.5 (2015): 623 - 627.
(project)

Activities

Invited Talk/Seminar Report
  • "Learning New Features for Effective Dense Matching Using MetaLevel Deep Neural Networks."
         Hong Kong University of Science and Technology, Oct 13, 2017. [Slides]
  • Reviewer/External Reviewer
    CVPR 2018, CVPR 2019, ICCV 2019
    IEEE Transactions on Image Processing
    IEEE Transactions on Circuits and Systems II (Signal Processing Section)
    IEEE Signal Processing Letters
    Teaching Assistant
    CSCI 3130 Formal Languages and Automata Theory. 2016-2017 first term
    ENGG 1110 Problem Solving by Programming. 2015-2016 second term
    CENG 2010 Digital Logic Design Laboratory. 2015-2016 second term
    CSCI 3130 Formal Languages and Automata Theory. 2015-2016 first term

    Honors & Awards

    Full Postgraduate Scholarship of Chinese University of Hong Kong, 2015-2017
    National First Prize in Contemporary Undergraduate Mathematical Contest in Modeling (1.4%), Nov 2013
    Merit Student of Jiangsu Province (Provincial-Level 0.06%), May 2013
    Meritorious Winner of the Mathematical Contest in Modeling (MCM), Apr 2013
    the National Scholarship (1%), Nov 2012
    the First Prize Scholarship (3%), 2011, 2012
    Merit Student (5%), 2011, 2012