Wenzheng Chen

I am a Ph.D. student at University of Toronto, advised by Prof. Sanja Filder and Prof. Kyros Kutulakos since 2017.

I interned at Algolux in 2018 summer, NVIDIA in 2018 fall and Snapchat in 2019 summer.

I obtained my master's degree at IRC, Shandong University, working with Prof. Yangyan Li, Prof. Changhe Tu and Prof. Baoquan Chen from 2014 to 2017. I obtained my bachelor's degree at Taishan College, Shandong University from 2010 to 2014.

Email  /  CV  /  Google Scholar  /  LinkedIn

profile photo

Generally, my research focuses on computational photography and 3D vision. More specifically, I work on predicting 3D properties from various imaging systems, including camera, renderer, structured light and SPAD. I am particularly interested in differentiating imaging systems and embedding them in deep learning, which takes advantages of both neural network features and all kinds of geometric, optical and physical knowledge brought by the imaging systems.

Structured Light
Auto-Tuning Structured Light by Optical Stochastic Gradient Descent
Wenzheng Chen*, Parsa Mirdehghan*, Sanja Fidler, Kyros Kutulakos (* Equal contribution)
CVPR, 2020  
paper / codes / bibtex

We present optical SGD, a computational imaging technique that allows an active depth imaging system to automatically discover optimal illuminations & decoding.

Optimal Structured Light a la Carte
Parsa Mirdehghan, Wenzheng Chen, Kyros Kutulakos
CVPR, 2018   (Spotlight Presentation)
paper / codes / bibtex

alacarte designs stuectured light patterns from a maching learning persepctive, where patterns are automatically optimized by minimizing the expected disparity matching loss.

Differentiable Render
Image GANs meet Differentiable Rendering for Inverse Graphics and Interpretable 3D Neural Rendering
Yuxuan Zhang*, Wenzheng Chen*, Jun Gao, Huan Ling, Yinan Zhang,
Antonio Torralba Sanja Fidler (* Equal contribution)
ICLR, 2021   (Oral Presentation)
arXiv / codes / bibtex

We explore StyleGAN as a multi-view image generator and train inverse graphcis from StyleGAN images. Once trained, the inver grahcis model further helps disentangle and manipulate StyleGAN latent code by feeding graphcis knowledge.

Learning Deformable Tetrahedral Meshes for 3D Reconstruction
Jun Gao, Wenzheng Chen, Tommy Xiang, Alec Jacobson, Morgan Mcguire, Sanja Fidler
NeurIPS, 2020  
arXiv / codes / bibtex

We predict deformable tetrahedral meshes from images or point clouds, which support arbitrary topologies. We also design a differentiable renderer for tetrahedron, allowing 3D reconstrucion from 2D supervison only.

Learning to Predict 3D Objects with an Interpolation-based Differentiable Renderer
Wenzheng Chen, Jun Gao*, Huan Ling*, Edward J. Smith*,
Jaakko Lehtinen, Alec Jacobson, Sanja Fidler (* Equal contribution)
NeurIPS, 2019  
arXiv / codes / bibtex

An interpolation-based 3D mesh differentiable renderer that supports vertex, vertex color, multiple lighting models, texture mapping and could be easily embedded in neural network.

Fast Interactive Object Annotation with Curve-GCN
Huan Ling*, Jun Gao*, Amlan Kar, Wenzheng Chen, Sanja Fidler (* Equal contribution)
CVPR, 2019  
arXiv / codes / bibtex

We predict object polygon contour from graph neural network, where a 2D differentiable rendering loss are introduced. It renders the polygon to segmentation mask and back propagate the loss to help optimize the polygon vertices.

Learned Feature Embeddings for Non-Line-of-Sight Imaging and Recognition
Wenzheng Chen*, Fangyin Wei*, Kyros Kutulakos,
Szymon Rusinkiewicz, Felix Heide (* Equal contribution)
SIGGRAPH Asia, 2020  
paper / codes / bibtex

We propose to learn feature embeddings for non-line-of-sight imaging and recognition by propagating features through physical modules.

Steady-state Non-Line-of-Sight Imaging
Wenzheng Chen, Simon Daneau, Fahim Mannan, Felix Heide
CVPR, 2019   (Oral Presentation)
arXiv / codes / bibtex

Estimating hidden objects from conventional images.

Earlier Work
Group optimization for multi-attribute visual embedding
Qiong Zeng*, Wenzheng Chen*, Zhuo Han, Mingyi Shi, Yanir Kleiman,
Daniel Cohen-Or, Baoquan Chen, Yangyan Li (* Equal contribution)
Visual Informatics, 2018

Multiple aspects embedding with bundle optimization.

Synthesizing Training Images for Boosting Human 3D Pose Estimation
Wenzheng Chen, Huan Wang, Yangyan Li, Hao Su, Zhenhua Wang,
Changhe Tu, Dani Lischinski, Daniel Cohen-Or, Baoquan Chen
3DV, 2016   (Oral Presentation)
arXiv / codes / bibtex

3D pose estimation from model trained with synthetic data and domain adaptation.

A Holistic Approach for Data-Driven Object Cutout
Huayong Xu, Yangyan Li, Wenzheng Chen, Dani Lischinski,
Daniel Cohen-Or, Baoquan Chen
ACCV, 2016

Instance segmentation with shapenet data augumentation.

Hallucinating Stereoscopy from a Single Image
Qiong Zeng, Wenzheng Chen, Huan Wang, Changhe Tu,
Daniel Cohen-Or, Dani Lischinski, Baoquan Chen
Eurographics, 2015

Creating hallucinated stereo map from a single image via depth layering.

I stole the website template from Jon Barron.