Chengzhou Tang

E2-480, 75 Chancellors Cir, Winnipeg, MB, R3T5V6, Canada

Email:firstname.lastname@umanitoba.ca

I lead the Spatial Intelligent System (SPINS) Lab at University of Manitoba (UM). Prior to UM, I was a Machine Learning Engineer at Apple and then a Research Scientist at Meta Reality Labs. I got my PhD degree from Simon Fraser University in 2020, supervised by Dr. Ping Tan. During my PhD, I had spent wonderful times as a research intern at Adobe Research with Dr. Oliver Wang and Microsoft Research with Dr. Lu Yuan.

My long-term research vision is to solve the `last-centimeter’ problem for autonomous intelligent system through an organic integration between visual perception and control. To this end, we have several recent short-term focuses: 1. simultaneous localization and mapping (SLAM) for dynamic scenes and systems, 2. learning efficient low-level computer vision (stereo, optical flow, grouping, etc) model with uncertainty, 3. learning visual representation of environment for model predictive control.

Prospective students: I have two fully funded positions for self-motivated PhD (preferred) and master students for 2025 Fall / 2026 Winter. If you are interested in working with me, please fill out this application form. Students from majors other than computer science are also encouraged to apply as long as you have a solid math or programming skill. Depending on the condition of each candidate, I am also open to other research topics relevant to the above focuses.

news

Sep 01, 2024	I have joined University of Manitoba as a tenure-track assistant professor.

selected publications

Vision

LSM: Learning Subspace Minimization for Low-level Vision

Chengzhou Tang, Lu Yuan, and Ping Tan

In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020

Bib PDF Code

@inproceedings{tang2020lsm,
  author = {Tang, Chengzhou and Yuan, Lu and Tan, Ping},
  title = {LSM: Learning Subspace Minimization for Low-level Vision},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2020},
}

Vision

BA-Net: Dense Bundle Adjustment Networks

Chengzhou Tang, and Ping Tan

In International Conference on Learning Representations (ICLR), 2019

Bib PDF Code

@inproceedings{tang2019banet,
  author = {Tang, Chengzhou and Tan, Ping},
  title = {BA-Net: Dense Bundle Adjustment Networks},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2019},
}

Graphics

Joint Direction and Stabilization for 360° Videos

Chengzhou Tang, Oliver Wang, Feng Liu, and Ping Tan

ACM Transactions on Graphics (TOG), 2019

Bib PDF

@article{tang2019tog,
  author = {Tang, Chengzhou and Wang, Oliver and Liu, Feng and Tan, Ping},
  title = {Joint Direction and Stabilization for 360° Videos},
  journal = {ACM Transactions on Graphics (TOG)},
  year = {2019},
}

Vision

GSLAM: Initialization-robust Monocular Visual SLAM via Global Structure-from-Motion

Chengzhou Tang, Oliver Wang, and Ping Tan

In International Conference on 3D Vision (3DV), 2017

Bib PDF Code

@inproceedings{tang2017gslam,
  author = {Tang, Chengzhou and Wang, Oliver and Tan, Ping},
  title = {GSLAM: Initialization-robust Monocular Visual SLAM via Global Structure-from-Motion},
  booktitle = {International Conference on 3D Vision (3DV)},
  year = {2017},
}