|
Zhouhui Lian(连宙辉)
|
Associate Professor, Ph.D. Wangxuan Institute of Computer Technology (WICT) Peking University No. 128 Zhongguancun North Street, Haidian District Beijing, China Phone: 86-10-82529245 Email: lianzhouhui@pku.edu.cn Our research group: IGCL |
News:
[2022-09-05] Our paper "SinMPI: Novel View Synthesis from a Single Image with Expanded Multiplane Images" has been accepted by SIGGRAPH Asia 2023 (source code and data are available)
[2023-03-08] I will serve as an Area Chair for NeurIPS 2023.
[2023-03-01] Our papers "DeepVecFont-v2: Exploiting Transformers to Synthesize Vector Fonts with Higher Quality" and "VecFontSDF: Learning to Reconstruct and Synthesize High-quality Vector Fonts via Signed Distance Functions" have been accepted by CVPR 2023 (source code and data are available)
[2022-09-14] I will serve as an Area Chair for CVPR 2023.
[2022-09-05] Our paper "DifferSketching: How Differently Do People Sketch 3D Objects?" has been accepted in ACM Transactions on Graphics (to be presented at SIGGRAPH Asia 2022)
[2022-07-07] Our paper "Semi-supervised Semantic Segmentation via Prototypical Contrastive Learning" has been accepted by ACM MM 2022 (source code and data)
[2022-04-26] Our paper "DCT-Net: Domain-calibrated Translation for Portrait Stylization" has been accepted in ACM Transactions on Graphics (to be presented at SIGGRAPH 2022) and our paper "CVFont: Synthesizing Chinese Vector Fonts via Deep Layout Inferring" has been accepted in Computer Graphics Forum (CGF) (Code and data will be available)
[2022-03-20] Our paper "Controllable Image Synthesis with Attribute-Decomposed GAN" has been accepted by IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) (Source code will be updated soon)
[2022-03-02] Our papers "Aesthetic Text Logo Synthesis via Content-aware Layout Inferring" and "Unpaired Cartoon Image Synthesis via Gated Cycle Mapping" have been accepted by CVPR 2022 (source code and data are available)
[2021-09-29] Our paper "CentripetalText: An Efficient Text Instance Representation for Scene Text Detection" has been accepted by NeurIPS 2021 (source code)
[2021-08-05] Our paper "DeepVecFont: Synthesizing High-quality Vector Fonts via Dual-modality Learning" has been accepted in ACM Transactions on Graphics (to be presented at SIGGRAPH Asia 2021)
Representative Works:
Canonical Forms for Non-rigid 3D Shape Retrieval (IJCV 2013, 3DOR 2017)
3D Shape Measures (rectilinearity, convexity) (IJCV 2010, CVPR 2012)
3D Shape Retrieval Contest (SHREC) (IJCV 2016, PR 2013(ESI Highly Cited Paper) , 3DOR 2015, 14, 11, 10)
Aesthetic Assessment and Font Recognition for Chinese Characters (CVPR 2017, IJCAI 2015)
Generation of Chinese Fonts (www.flexifont.com, SIGGRAPH (TOG) 19, 20, SIGGRAPH ASIA (TOG) 19)
Texture Transfer and Text Effects Transfer (CVPR 18, 19, 20)
Bio:
I was graduated from School of Automation Science and Electrical Engineering, Beihang University in Sep. 2011 with Ph.D. From Jan. 2008 to Jan. 2009, supported by the China Scholarship Council (CSC), I visited School of Computer Science, Cardiff University, UK as a visiting scholar. Then, I worked as a Guest Researcher at the Information Technology Lab (ITL), National Institute of Standards and Technology (NIST), U.S. from Jul. 2009 to Jul. 2011. After that, I came back to China and worked as a postdoctoral researcher at ICST, Peking University, where I became an Assistant Professor in Jun. 2013 and was promoted to Associate Professor in Aug. 2015.
My research interests include Computer Graphics, Computer Vision, Pattern Recognition and Artificial Intelligence. More specifically, we have been working on the following topics in last few years. (1) Easy/automatic generation of large-scale Chinese font libraries. we proposed a series of solutions including radical assembling based and style learning based methods, making the easy/automatic generation of large-scale Chinese fonts become possible. (2) Calligraphic aesthetic assessment and font recognition. We proposed to adopt several machine learning approaches (e.g., artificial neural networks, sparse coding, etc.) to effectively evaluate the aesthetic visual quality of Chinese handwritings and recognize font styles. (3) Non-rigid 3D shape retrieval. We developed a number of canonical form construction methods to covert non-rigid 3D shape searching tasks into rigid cases, for which several view-based 3D shape matching approaches are proposed to effectively address the problems.
Students:
PhD Students:
Master Students:
Shilong Li
Bojun Xiong
Hua Li
Jialei Cui
Yuqing Wang
Alumni
Zeqing Xia (PhD 2023)
Yizhi Wang (PhD 2022)
Zenggui Chen (Master 2023)
Jie Chen (Master 2021)
Yue Gao (Master 2021)
Shuseng Tang (Master 2021)
Tao Sheng (Master 2021)
Xiang Zhang (Master 2021)
Yifang Men (Master 2020)
Xiao Sun (Master 2020)
Yichen Gao (Master 2020)
Wenguan Wang (Master 2020)
Juncheng Liu (PhD 2019)
Yuan Guo (Master 2019)
Yue Jiang (Master 2019)
Xudong Chen (Master 2017)
Chendong Liu (Master 2017)
Weikang Song (Master 2016)
Rongju Sun (Master 2016)
Wanqiong Pan (Master 2015)
Hao Sun (Master 2015)
Jun Zhang (Master 2015)
Wenyan Dong (Master 2015)
Tianyang Yi (Master 2014)
Chenxi Wang (Master 2013)
Yang Liu (Master 2012)
Hongmei Li (Master 2012)
Reading List:
I have published more than 70 academic papers, including 1 ESI Highly Cited Paper (top 1%), 5 full papers in ACM Transactions on Graphics (TOG), 1 full paper in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 3 full papers in International Journal of Computer Vision (IJCV, a top-tier journal in Computer Vision), and several top conference (e.g., SIGGRAPH/SIGGRAPH Asia, CVPR, NeurIPS, IJCAI, AAAI, MM, etc.) papers. I was the principal investigator of several projects funded by the National Natural Science Foundation of China, Beijing Nova Program of Science and Technology, etc., and have led our group to successfully develop two practical systems for the generation of Chinese fonts. These systems and techniques related to computer fonts have been adopted by a company called Founder Handwriting since 2016 to create font products that have served millions of users, making significant social and economic benefits. Our works on 3D shape analysis and retrieval have been cited more than 2000 times (until March 2022) by papers published in prestigious journals and conferences including IJCV, TPAMI, TOG, CVPR, etc.
Publications (selected): Back to top
Spring (2014~now): Elementary Number Theory
Fall (2017~now): Intelligent Optimization Methods
2021, China Patent Award of Excellence (ranked 1st) 2020-2021, CVPR Outstanding Reviewer 2019, Wu Wenjun Outstanding Youth Award of Artificial Intelligence, CAAI 2018, First class award of "Wang Xuan young teacher award" in ICST, Peking University 2014, First class award of the 14th teaching skill contest for young teachers in Peking University (top 2) 2012, Excellent postdoctoral researcher in Peking University (top 1%)
Area Chairs for NeurIPS-23, CVPR-23, ICCV-21; SPCs for IJCAI-21, AAAI-22. IEEE/ACM/CAAI member, CCF/CSIG senior member Reviewers for journals including TPAMI, ACM Transactions on Graphics, TIP, TMM, ACM TOMM, Pattern Recognition, IJDAR, Multimedia Systems, Parallel Computing, Computers & Graphics, The Visual Computer, International Journal of Machine Learning and Cybernetics, Displays, etc. PC members for CVPR, ICCV, IJCAI, AAAI, MM, ECCV, ICDAR, ICIP, ACCV, WACV, EG 3DOR, etc. Talk: 2020-08-02, VALSE 2020, "Glyph Image and Shape Synthesis and Their Applications" Talk: 2020-05-18, CVPR pre-conference 2020, "Controllable Person Image Synthesis with Attribute-Decomposed GAN"
Last update on Aug. 13, 2020
visits since May. 2015