Scan contact card
I am a master student in Electrical and Computer Engineering at Johns Hopkins University, advised by Dr. Jerry Prince. I am currently working as a research assistant at the Image Analysis and Communications Laboratory (IACL). My research interests are computer vision, medical image analysis and machine learning.
Before that, I received my B.E. at Southeast University.
I was working as an intern at Philips North America in 2018 summer.
[View my Resume]
This section contains past projects I worked on.
This is the current project I am working on. My ultimate task is to develop better segmentation and analysis methods for thalamic & sub-thalamic nuclei based on multimodal MRI such as DTI/DKI and MPnRAGE. I have been conducting some experiments including investigating several fiber tractography methods solving crossing fiber problem, visualizing multimodal data using TSNE and UMAP, investigating optimization tools for difffusion kurtosis claculation and the cause of error estimation. Read more..
This project was to estimate the 6 degrees-of-freedom pose of the mobile C-arm imaging device based on a single image. The motivation was to determine the best device position within limited X-ray shots in order to reduce unnecessary radiation exposure administered by the system. My work involved investigating different design markers by which to encode the device pose, and I decoding the pose through segmentation and registration methods. In addition, I implemented an end-to-end learning scheme to train a deep neural network based on the whole image information, using tens of thousands synthetic X-ray data I generated. Read more..
Practicing handwriting using the copybook of calligraphy is suffering because we usually don't know how good we write and how much we can improve it. So we developed Master Writing to help people conquer it. Master Writing is an IOS application that helps people improve handwriting by showing how a master is writing and let people imitate it. Users can learn how to master different fonts by practicing letters, words and sentences. The strokes, force and angel are recorded when users are writing. These information are compared with a master's, and various feedbacks are provided interactively at the same time to tell people how to improve. This app was developed by my two friends and me within 36 hours during Hophacks 2018 spring and won Second Price Overall. Read more..
This is my course project for Wavelets course. The motivition of this project is to utilize wavelets characteristics to address the privacy concern in medical image database system. During the course I was thinking if we can store sensitive patient information invisibly using wavelets watermarking technique. I designed my soluttion with the several techniques to enhance the security and robustness in data storing and transfering. Read more..
Gas-inhalation MRI is a novel imaging technique to measure multiple brain hemodynamic parameters. This technique uses two physiological measures, specifically arterial CO2 and O2 time course, as input and BOLD MRI signal time course as output, and employs a linear model to determine the association between gas challenge and MRI signal, which is related to vascular properties of the brain. A prerequisite of this analysis is that the CO2 and O2 time course needs to be aligned temporally. In this project, I developed a simple and fully automated method to synchronize the CO2 and O2 signals, based on time-shifted correlation of these two signals. Read more..
This is my bachelor graduation project in 2016. Inspired by the paper "Grid-cell representations in mental simulation" published in August 2016, I was thinking if grid-like representations can be activated given more abstract and simple cues, and can be detected using fMRI. The motivation was to simplify the original experimental paradigm and make it possible to use in wider clinical trails and measurements. I first implemented the navigation task in virtual-reality city using Unity3D, repeated the results in the paper. Then based on my assumption, I designed a modified paradigm using PyQt, which only kept the location relationship in previous task and represented them in 2D plane, without other 3D visual cues. I was awarded Outstanding Graduation Project in June 2017. Read more..
This project was to investigate super resolution reconstruction of lateral ventricle volume from fetal brain MRI data. The most challenging aspect of the project was to deal with the inter-slice noise caused by fetal head motion. We implemented a reconstruction method in my co-authored paper published in ISMRM, where we first modeled degradation and realigned low-resolution images via convex optimization in order to reduce motion artifacts. This result along with additional data from the same gestation week was used in a sparse reconstruction algorithm to achieve the desired super-resolution. Read more..
This is a small prject I have been working on when I first joined Medical Electronics Laboratory in Southeast University. My task was to extract and analyze the real-time heart rate from PPG signals. I investigated Hilbert-Huang transform (HHT), short-time Fourier transform (STFT) and continuous wavelets transform (CWT) in this project. Read more..
This student project was conducted on 2016 at my junior year. It stemmed from one of my thoughts: can we detect lip movements with micro-camera, transfer lip language to synthetic voice, thus enabling the mute to communicate just like us? we took 10000 photos of Chinese speakers speaking basic vowels, conducted automatic detection of faces and lips with OpenCV, and trained a neural network to recognize basic Chinese vowels. We built a lip language reader using Raspberry Pi and a camera with 92.03 recognition accuracy. This project was also funded by National Undergraduate Innovation & Entrepreneurship program. Read more..
I worked on this student project with two friends when we were sophomores. We built an reliable aircraft monitoring system with 200km detection zone, based on ADS-B technology. We improved the receiver by making an extensional antenna ourself. We obtained aircraft data using rtl-sdr and provided the website service for two years to visualize the real-time aircraft information. We also conducted blind sourse seperation experiments on aliased 1090ES signals. We won about $3000 fund from National Undergraduate Innovation & Entrepreneurship program, and was awarded Outstanding Project. Read more..
A collection of articles, conference papers or presentations.
IACL, Johns Hopkins University, January 2018 - Present
Neuroimaging Department,Nanjing Drum Tower Hospital, June 2016 - June 2017
Compressed Sensing 19 Spring, Machine Learning 18 Fall at Johns Hopkins University
Programming and Algorithmic Language 14 Fall & 15 Spring at Southeast University
Research Intern at Image Guided Surgery group, May 2018 - Aug 2018
Software Engineer Intern at Donglan Digital Co., Ltd, July 2014 - August 2014
This section contains my posts.