Zhengxia Zou, Ph.D.


"Keep Calm and Don't Overfit."


I am currently a postdoc research fellow at the University of Michigan, Ann Arbor. I got my B.S. degree and Ph.D. degree from Beihang University in 2013 and 2018. My research interests include computer vision and related applications in remote sensing, self-driving vehicles, and video games. ([Google Scholar Profile], [Github Profile], Email: zzhengxi [at] umich [dot] edu.)

NEW (02/2021): One paper accepted to CVPR 2021 as an oral presentation ( Preprint | Project | GitHub ).

NEW (01/2021): Our Stylized Neural Painting was launched as a new feature in Remini - an AI photo enhancer with over 50M users worldwide. ( Google Play | IOS App Store )

NEW (01/2021): One paper accepted to IEEE Transactions on Image Processing ( Preprint | Github ).

NEW (01/2021): Check out our new work on inverse autostereograms ( Preprint | Project | GitHub ).

NEW (12/2020): I will be serving as a Senior Program Committee (SPC) member in IJCAI 2021.

NEW (12/2020): One paper accepted to AAAI 2021.

NEW (10/2020): Check out our preprint paper on video sky blending and harmonization ( Preprint | Project | GitHub ).

NEW (09/2020): Our neural rendering paper was accepted by TPAMI. A great pleasure working with NetEase Fuxi Lab, congratulations!

NEW (07/2020): One paper accepted to ACM MM 2020.

NEW (02/2020): One paper accepted to CVPR 2020.

Selected publications and preprints from 2019 to 2021 (*For my full publication list, please visit my google scholar profile)

Zhengxia Zou, Zhenwei Shi, Yuhong Guo, and Jieping Ye. Object Detection in 20 Years: A Survey. CoRR, abs/1905.05055, 2019. [PDF]

This paper extensively reviews 400+ papers of object detection in the light of its technical evolution, spanning over a quarter-century’s time (from the 1990s to 2019). A number of topics are covered in this paper, including milestone detectors in history, detection datasets, metrics, fundamental building blocks of detection systems, speed up techniques, and the recent state of the art detection methods.

- High-impact Citations: Stanford CS231n: Convolutional Neural Networks for Visual Recognition 2019, 2020 | GeorgiaTech-GT8803: Data analytics using deep learning 2019 (Fall) 

- Media Coverage:   密歇根大学40页最新论文带你全面了解目标检测 | 专栏连载: 二十年的目标检测大综述专栏文章: 2019-2020年必看的十篇「深度学习领域综述」论文  大盘点: 2019年4篇目标检测算法最佳综述

- In Other Languages:  English-to-Chinese (1)English-to-Chinese (2) 





Zhengxia Zou, Tianyang Shi, Yi Yuan, and Zhenwei Shi. NeuralMagicEye: Learning to See and Understand the Scene Behind an Autostereogram. CoRR, abs/2012.15692, 2020. [PDF] [1min-DemoVideo] [Project] [Github] [Google Colab]

 - Media Coverage:      密西根大学网易北航搞了个奇怪研究:多数人类看不懂的立体图,AI现在看懂了



Zhengxia Zou, Sen Lei, Tianyang Shi, Zhenwei Shi, and Jieping Ye. Deep Adversarial Decomposition: A Unified Framework for Separating Superimposed Images. IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2020. [PDF] [1min-DemoVideo] [Github]


Zhengxia Zou, Tianyang Shi, Zhenwei Shi, and Jieping Ye. Adversarial Training for Solving Inverse Problems in Image Processing. IEEE Transactions on Image Processing, 2021, in press. [PDF] [1min-DemoVideo] [Github]


Zhengxia Zou, Wenyuan Li, Tianyang Shi, Zhenwei Shi, and Jieping Ye. Generative Adversarial Training for Weakly Supervised Cloud Matting. IEEE International Conference on Computer Vision (ICCV) 2019. [PDF] [Github]


Yinglin Duan*, Tianyang Shi*, Zhengxia Zou* (*Equal contribution), Yenan Lin, Zhehui Qian, Bohan Zhang, and Yi Yuan. Single-Shot Motion Completion with Transformer. CoRR, abs/2103.00776, 2021. [PDF]


Yinglin Duan*, Tianyang Shi* (*Equal contribution), Zhengxia Zou, Jia Qin, Yifei Zhao, Yi Yuan, Jie Hou, Xiang Wen, and Changjie Fan. Semi-Supervised Learning for In-Game Expert-Level Music-to-Dance Translation. CoRR, abs/2009.12763, 2020. [PDF]


Tianyang Shi, Zhengxia Zou, Xinhui Song, Zheng Song, Changjian Gu, Yi Yuan, and Changjie Fan. Neutral Face Game Character Auto-Creation via PokerFace-GAN. ACM International Conference on Multimedia (ACM MM) 2020. [PDF] [Github]


Tianyang Shi, Zhengxia Zou, Yi Yuan, and Changjie Fan. Fast and Robust Face-to-Parameter Translation for Game Character Auto-creation. AAAI Conference on Artificial Intelligence (AAAI) 2020. [PDF]


Wenyuan Li, Zhengxia Zou, and Zhenwei Shi. Deep Matting for Cloud Detection in Remote Sensing Images. IEEE Transactions on Geoscience and Remote Sensing, in press, 2020.  [PDF]


Zhengxia Zou*, Tianyang Shi* (*equal contribution), Wenyuan Li, Zhou Zhang, and Zhenwei Shi. Do Game Data Generalize Well for Remote Sensing Image Segmentation? Remote Sensing, in press, 2020. [PDF]

[GTA-V-SID] - A synthetic remote sensing image dataset from the video game GTA-V for semantic segmentation.

Sen Lei, Zhenwei Shi, and Zhengxia Zou. Coupled Adversarial Training for Remote Sensing Image Super-resolution. IEEE Transactions on Geoscience and Remote Sensing. 58(5), May 2020, pp. 3633 - 3643. [PDF]


Tianyang Shi, Yi Yuan, Changjie Fan, Zhengxia Zou, Zhenwei Shi, and Yong Liu. Face-to-Parameter Translation for Game Character Auto-Creation. IEEE International Conference on Computer Vision (ICCV) 2019. [PDF]


Guest Editor
Peer-Review Services
  • Senior Program Committee (SPC) member of IJCAI 2021.
  • NeurIPS 2019, 2020 PC Member; ICLR 2021 Reviewer; AAAI 2020, 2021 PC Member; CVPR 2020 PC Member; ACCV 2020 PC Member; WACV 2021 PC Member.
  • IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Transactions on Image Processing; IEEE Signal Processing Magazine; IEEE Transactions on Geoscience and Remote Sensing; IEEE Signal Processing Letters; IEEE Geoscience and Remote Sensing Letters; Remote Sensing; Infrared Physics and Technology; International Journal of Remote Sensing; Electronics; Journal of Computational Methods in Sciences and Engineering; Journal of Marine Science and Technology; Computational Intelligence and Neuroscience; Journal of Spectroscopy; The Visual Computer; International Journal of Machine Learning and Cybernetics.
  • Best Reviewer of Infrared Physics & Technology (2017).
Invited Talks
  • Art-centric Neural Rendering and Augmented Reality. University of Texas, Austin. Dec 2020.
  • Neural Rendering for Visual Editing and Visual Arts. Ocean University of China. Dec 2020.
  • Object Detection in 20 Years. Didi Chuxing, Beijing, China. Jul 2019.
  • Adversarial Training for Solving Inverse Problems. Didi Chuxing, Beijing, China. Jul 2019.
  • Introduction to Deep Learning and Computer Vision - Past, Present and Future. M-Culture Salon, University of Michigan, Ann Arbor. Jan 2020.
Summer Exchange
  • Outstanding PhD dissertation award (nomination) of CSIG. Nov. 2019 (中国图象图形学会优博论文提名奖)
  • Outstanding PhD dissertation award of Beihang University. Sept. 2019 (北航优博论文奖)
  • Top-10 Graduate Students" of Beihang University (top-10/4,000). Jun. 2018 (北航十佳博士研究生)
  • Outstanding Graduates of Beijing. Jul. 2018 (北京市优秀毕业生)
  • Best Reviewers of the "Infrared Physics & Technology", 2017 (2017年度最佳审稿人)
  • Ph.d. National Scholarship (top-100/4,000). Dec. 2016 (研究生国家奖学金)
  • Ph.d. Excellent Scholarship of Beihang University (top-100/4,000). Dec. 2017 (北航卓越基金)
  • Grand prize of the "Baogang Scholarship" (50 students selected in China, ranking 1st in Beihang University). Nov. 2012 (宝钢奖学金特等奖,2012年全国仅50名)
  • First prize of the "Beijing College Student Physics Competition" (Top 50 in Beijing, ranking 1st in Beihang University). Dec. 2010 (北京市大学生物理竞赛一等奖,校第一名)
  • First prize of the "Beihang College Student Mathematics Competition" (top-30/1,000). Jun. 2010 (北航大学生数学竞赛一等奖)
  • First prize of the "Fengru Cup Scientific Competition" (top-20/1,000). May. 2011 (北航冯如杯一等奖,第一作者,获研究生推免资格)
  • Second prize of the "Fengru Cup Scientific Competition". May. 2012 (北航冯如杯二等奖)
  • Second prize of the "Beijing College Student Scientific Competition". Jun. 2011 (首都挑战杯二等奖,第一作者)
  • Third prize of the "National College Student Mathematics Competition". Jun. 2010 (全国大学生数学竞赛三等奖)
  • Second prize of the "Beihang College Student Physics Competition". May. 2010 (北航大学生物理竞赛二等奖)
  • Third prize of the "Zhou Peiyuan Mechanics Competition for College Students". May. 2011 (全国周培源大学生力学竞赛三等奖,所在非力学专业唯一晋级选手)
  • Bronze Award of the "2nd Beihang University Student Art Show: Instrumental Competition" (北航大学生艺术展演——器乐专场比赛铜奖)
Robot Worm

This robot worm was made during my sophomore years. Me and my group members spent 3 months and made it from a pile of metal scraps. We had an amazing time working together. We finally won the first prize in a student contest of our school and beat all the senior teams.
Visualization of Optimization Algorithms

Numerical optimization is one of the core math foundations in image processing and machine learning. But I remember in the beginning of my Ph.D. years, the math behind always made me frustrated 🙁🙁.

During the winter vacation of 2016, I decided to make a change. I revisited some well-known optimization methods (e.g., Gradient Descent, Newton/Quasi-Newton Method, ALM, etc.), and made a series of GIF visualizations to show how these algorithms behave dynamically. Check out this blog for more details and I hope it will help you better understand these algorithms.


Copyright (C) 2020 By Zhengxia Zou.