Xiaochen Zhou

I am a graduate student at Washington University in St. Louis, where I major in computer vision, image processing and machine learning. I am a research assistant in Vision & Learning Group, advised by Ayan Chakrabarti.

I defensed my bachelor degree at Beihang University, and achieved an amazing internship experience in Megvii Face++, working on vehicle re-identification.

I am now looking for the chance for further PhD study in computer vision, image processing and machine learning. I would love to make contributions to entertainment or art industry owing to my crazy interests on video game, sci-fi movie and oil painting. Working for one's interests is extremely awesome.

Email  /  CV  /  Research Interests  /  LinkedIn  /  Github

profile photo
Research

I'm interested in computer vision, machine learning, optimization, and image processing. More specifially, I'm interested in 3D reconstruction and image texuture synthesis.

Contour-to-Image Reconstruction through Neural Network
Implementation practice of paper Smart, Sparse Contours to Represent and Edit Images, 2018

In this project, a generative adversial network is trained to generate and edit image through contour domain, where canny edge detection is used to extract contour information through original input.

Github
Canny Edge Detector
Algorithms Implementation, 2019

Implemented the traditional algorithm with Python and Scipy. This algorithm will be used in edge detector task, and will be implemented in tensorflow version for contour2img task.

Partial Harmonization Network
Implementation practice of paper Deep Painterly Harmonization, 2019

Training a neural network on style transferring and optimizing the coarse general output. This network allows you to copy and paste some part of image element from other image to your target image and smooth the artifacts

Github
Single-view 3D Model Reconstruction
Course Project in Computer Vision, 2018

Computing camera calibration parameters with vanishing points and reconstructing the object through search algorithm. This project is my first step to single view 3D reconstruction. Now I am playing with PointNet and point cloud SIFT algorithm.

Learning Discriminative 3D Shape Representations by View Discerning Networks
Biao Leng, Cheng Zhang, Xiaochen Zhou, Cheng Xu, Kai Xu
TVCG, 2018  

Two score units are devised to evaluate the quality of each projected image with score vectors.

Improved Panoramic Representation via Bidirectional Recurrent View Aggregation for 3D Model Retrieval
Cheng Xu, Cheng Zhang, Xiaochen Zhou, Biao Leng
IEEE Computer Graphic and Application, 2018  

A novel deep neural network, recurrent panorama network (RePanoNet), is designed to extract features in a panoramic view, encouraging the network to recognize the original 3D shape.

Emphasizing 3D Properties in Recurrent Multi-View Aggregation for 3D Shape Retrieval
Cheng Xu, Biao Leng, Cheng Zhang, Xiaochen Zhou
AAAI, 2018  

We designed an encoder-decoder recurrent feature aggregation network (ERFA-Net) to emphasize the 3D properties of 3D shapes in multi-view features aggregation.

Service
TA on CSE559A: Computer Vision - Fall 2019

Great thank for the website template!