Zhengyang Shen
|
I am a research scientist at Google AR Team. I graduated from UNC-Chapel Hill, working on 3d computer vision, geometry processing and medical image analysis with Prof. Marc Niethammer. Before coming to UNC, I earned my bachelor's degree from Beihang University, where I worked on machine learning advised by Prof. Yan Xu at MSRA. I worked on three packages Mermaid, EasyReg and ShapMagn. Mermaid is a image registration toolkit making use of automatic differentiation for various of fluid-based registration models. EasyReg supports a wide range of optimization/learning based registration, data augmentation and segmentation approaches. A general introduction for image registration approaches can be found here. ShapMagn is a point cloud processing package supporting a wide range of optimization/learning based registration approaches, data augmentation pipeline and landmark detection approach. Demos are included in these repositories.
Journal Reviewer: TPAMI, TIP, TMI, MIA. Email: zyshen@cs.unc.edu |
News
- Sep 2021 » Our paper on complicated large-scale point cloud registration is accepted at NeurIPS 2021.
- Sep 2021 » Our research toolkit for large scaling point cloud registration is released.
- Sep 2021 » I will work as a part-time student researcher at Facebook Reality Lab this fall for facial geometry processing.
- May 2021 » I will do research intern at Facebook Reality Lab this summer for facial geometry processing.
- Jul 2020 » One paper accepted at ECCV and one at MICCAI.
- May 2020 » I will do research intern at Google this summer for deepfake video detection.
- Nov 2019 » Our scientific package for fluid based image registration is released.
- Sep 2019 » One paper accepted at NeurIPS 2019.
- Jun 2019 » I will do research intern at Siemens Healthnieers this summer for Brain Hemorrhage detection.
- Jun 2019 » Our work on extension of classic LDDMM is on arxiv.
- Mar 2019 » One paper accepted at CVPR 2019.
- Aug 2017 » I joined UNC CS as a PhD student.
Conference
|
Accurate Point Cloud Registration with Robust Optimal Transport
NeurIPS 2021
|
|
|
|
|
|
|
|
Networks for Joint Affine and Non-parametric Image Registration
CVPR 2019
|
|
Journal
|
Learning Multi-level Feature For Human Action Recognition
Pervasive and Mobile Computing, 40(9), 2017
|
|
Parallel Multiple Instance Learning for Extremely Large Histopathology Image Analysis
BMC Bioinformatics, 18(1), 2017
|
|
End-to-End Subtitle Detection and Recognition for Videos in East Asian Languages via CNN Ensemble with Near-Human-Level Performance
Signal Processing: Image Communication, 60(10), 2016
|
Miscs
- My Chinese name is 沈正阳.
- I am from Songjiang, Shanghai, a beautiful place well known as 云间.
- Thanks Jie Lei's templete :)