Zhexiong Wan's Homepage

About Me

I am currently a Ph.D. candidate in Information Engineering, supervised by Prof. Yuchao Dai in NPU-CVR (website in Chinese) lab at Northwestern Polytechnical University (NPU), Xi’an, China. I am currently a visiting PhD student at National University of Singapore (NUS) under the supervision of Prof. Gim Hee Lee.

Before that, I received B.E. degree in Computer Science at NPU in 2019, during which I participated in several student competitions and received the Outstanding Undergraduate Student and Outstanding Undergraduate Thesis awards.

My research lies at the intersection of Low-Level Vision and 3D Vision, with a special focus on Event-based vision. My research interests include motion estimation (optical flow and scene flow estimation), motion segmentation, motion deblurring, video frame generation (interpolation, reconstruction and prediction), etc.

I expect to be on the job market in 2025. I would appreciate a ping if you see a position I might fit.

News

[Dec. 2023] One paper about video frame prediction with event camera is accepted to AAAI 2024. Congratulations to the collaborators!
[Dec. 2023] We won the 2nd place of VFI track in the 2023 The Greater Bay Area International Algorithm Case Competition . Congratulations to the collaborators!
[Sep. 2023] One paper about parametric optical flow is accepted to NeurIPS 2023. Congratulations to the collaborators!
[Jul. 2023] One paper about scene flow estimation with event camera is accepted to ICCV 2023. Thanks to the collaborators!
Show more

Publications [Google Scholar][DBLP]

AAAI

Video Frame Prediction from a Single Image and Events

Juanjuan Zhu*, Zhexiong Wan*, Yuchao Dai (* equal contribution)

AAAI Conference on Artificial Intelligence (AAAI), 2024.

PDF Code Project Page BibTex Waiting for updates
NeurIPS

Continuous Parametric Optical Flow

Jianqin Luo*, Zhexiong Wan*, Yuxin Mao, Bo Li, Yuchao Dai (* equal contribution)

Advances in Neural Information Processing Systems (NeurIPS), 2023.

PDF Code Project Page BibTex
ICCV

RPEFlow: Multimodal Fusion of RGB-PointCloud-Event for Joint Optical Flow and Scene Flow Estimation

Zhexiong Wan, Yuxin Mao, Jing Zhang, Yuchao Dai

IEEE International Conference on Computer Vision (ICCV), 2023.

PDF Supp Code Project Page BibTex
CVPR

Joint Appearance and Motion Learning for Efficient Rolling Shutter Correction

Bin Fan*, Yuxin Mao*, Yuchao Dai, Zhexiong Wan, Qi Liu

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023.

PDF Supp Code Video BibTex
TIP

Learning Dense and Continuous Optical Flow from an Event Camera

Zhexiong Wan, Yuchao Dai, Yuxin Mao

IEEE Transactions on Image Processing (TIP), 2022.

PDF Supp. Code Project Page BibTex
TCSVT

Deep Idempotent Network for Efficient Single Image Blind Deblurring

Yuxin Mao*, Zhexiong Wan*, Yuchao Dai, Xin Yu (* equal contribution)

IEEE Transactions on Circuits and Systems for Video Technology (TCSVT), 2022.

PDF Code Project Page BibTex
Generative Transformer for Accurate and Reliable Salient Object Detection

Yuxin Mao, Jing Zhang, Zhexiong Wan, Yuchao Dai, Aixuan Li, Yunqiu Lv, Xinyu Tian, Deng-Ping Fan, Nick Barnes

arXiv 2021.

PDF Code BibTex
arXiv

PRAFlow_RVC: Pyramid Recurrent All-Pairs Field Transforms for Optical Flow Estimation in Robust Vision Challenge

Zhexiong Wan, Yuxin Mao, Yuchao Dai

ECCV workshop, Robust Vision Challenge 2020.

PDF Live Leaderboard BibTex 2nd place of Optical Flow track

Project Experience

3D Dynamic Perception using Neuromorphic Vision Sensor Jan. 2021 -- Dec. 2022

Student Leader, supervised by mentor
Introducing event cameras to address the challenge of 3D dynamic perception in real scenarios with high dynamics.
This project is sponsored by Zhejiang Lab Open Research Funding.

SAR-Optical Image Matching Oct. 2020 -- Jun. 2021

Major contributor to won the championship in a Chinese AI Challenge.
The task is to match two heterogeneous source images to obtain the position of a small optical image block in a large synthetic aperture radar (SAR) image.
Our proposed network uses an attention mechanism and fast large kernel Fourier convolution and achieves first place in terms of speed and accuracy.

Undergraduate Robotics Maker Team at NPU Sept. 2016 -- Jun. 2019

Team Leader since Sept. 2017.
Our team have won many awards in several robot competition in China. My responsibility is to deploy visual recognition algorithms for robots, including the following applications:
Enemy identification (up to 200~FPS) and automatic attacking for the multi robot fierce confrontation competition, RoboMaster.
Identifying danger signs and trapped people for environment perception and mapping for urban search and rescue robot competition, RoboCupRescue China Open. CGTN Interview
Indoor positioning and navigation of UAVs using visual markers or gestures. Partial Code
Lead team members in collaboration and make project plans to achieve all goals.

Selected Honors and Awards

2nd place of VFI track in the 2023 The Greater Bay Area International Algorithm Case Competition	2023
The Innovation Foundation for Doctor Dissertation of NPU	2023
First-Class Academic Scholarships for PhD Students in NPU	2020&2022&2023
2nd place of Optical Flow track in the ECCV 2020 Robust Vision Challenge	2020
First-Class & Outstanding Advisor of RoboMaster University Championship	2019
Championships of RoboCup@Rescue China Open & China Robot Competition	2018
First-Class of The 11th Chinese College Students Computer Design Contest	2018
First-Class & Application Innovation of ASC17 Student Supercomputer Challenge	2017

Additional Information

Review services: TPAMI, CVPR, ICCV, ECCV, ICML, NeurIPS, ICLR, MM, ACCV, ICRA, IROS, etc.
Skills: Python, C/C++, Matlab, etc. Maintain the lab cluster with 20+ nodes and 160+ GPUs. GPU Monitor
Hobbies: Electronics, music, animation.

Last Update: