RUIZHI HONG
Objective
Developer
* ***** ********** ** ****** programming
* ***** ********** ** **** Modeling with contest Awards
Education
Aug, 2008~Present
Master of Science, University of Missouri-Columbia,
Electrical and Computer Engineering Department
3D reconstruction based on Multi-view images
Robot path planning implemented on parallel device
Aug, 2004~Jun, 2008
B.S in Electro-Optical Information Engineering,
Harbin Institute of Technology(HIT), China
Publications
D. Lam, R.Z. Hong, G.N Desouza, “3D Human Modeling Using Virtual Multi-View Stereopsis and Object-Camera Motion Estimation”, IEEE IROS 2009
R.Z. Hong, G.N Desouza, “A Real-Time Path Planner for a Smart Wheelchair Using Harmonic Potentials and a Rubber Band Model”, IEEE IROS 2010 (accepted)
Skills
Programming Tool: GCC, Visual Studio, FLTK, SAS, Auto-CAD
Libraries: CUDA, OpenGL, OpenCV, Aria, ARTookit
Programming Language: C, MATLAB,
Other skills: Mathematical Modeling, Image Processing
Research Experience
March, 2009~Present
Implemented robot path planning based on harmonic hield and optimization algorithm with CUDA
Implemented harmonic potential algorithm to get the original path
Built an elastic optimization model and got the optimized path
Implemented algorithms on CUDA divide in order to achieve real-time performance. Manipulated Arial Robot movement on simulator and real robot. Collected laser sensor data to update map information simultaneously and moved the robot in real time.
Build a optimization model combine rubber band and potential field together in order to achieve a smooth path
Implement kalman filter for object tracking and prediction. And make the robot avoid dynamic obstacles efficiently.
Aug, 2008~Jun, 2009
3D reconstruction based on multi-view virtual cameras
Used SIFT algorithm to detect feature points and then recontructed 3D position of features from stereo images
Assuming a sequence of images were taken by large amount of virtual cameras at the same time, match common features from different pairs of stereo images and calculate virtual camera positions based on common features
Reconstructed 3D model using Patch-based Multi-view Stereo Software(PMVS) with virtual cameras data input
Course Related Projects
Feb, 2010~May, 2010
Puma 260 control game based on vision
Implement kinematic and inverse kinematic control robot arms
Use MATLAB image processing to detect objects by camera vision and design attract and defense strategy
Oct, 2008~Dec, 2008
OpenGL 3D ‘Bombman’ game
Render 3D scenes by OpenGL and control elements by user interface
Organize AI control and design player functions of the game
Mar, 2009~May, 2009
Sparse model face recognition
MATLAB 2D image processing for background modeling
Build sparse face recognition face recognition model and test with different database
Mathematical Modeling
Feb, 2007
Organ(Kidney) Donation Center Planning Strategy (MATLAB, SAS)
The Interdisciplinary Contest of Modeling (USA)
Analyze human age group distribution, including expected life time with or without kidney change. Built possibility model in SAS.
Simulate data of human key DNA pairs for kidney (6 pairs), considering matches performance, estimate the opportunity cost and risk in MATLAB.
Implement Analytic Hierarchy Process (AHP) model considering parameters such as waiting time, age, match performance, privilege (kids priority), potential risk, and build a scoring system and schedule kidney exchange for a local donation center.
Planning promotion policy to improve human donation motivation, such as donor family members’ privilege in queue system, and also consider multiple centers cooperate.
Summary and analyze final performance of the whole donation system in MATLAB and SAS
Oct, 2007
Beijing Olympic Public Traffic System design (MATLAB, SAS)
The Mathematical Contest of Modeling (China)
Analyze official data of Beijing transportation system.
Fit curves total running time of distance, exchange time, subway exchange, local time (rush hour or not), and get approximate relationship function in SAS.
Design searching algorithm in order to get shortest time routine and shortest exchange time routine in MATLAB.
Bring the idea of creating possible new transportation routines, including building new subway, simulate how much performance to improve the running time, and reduce the crowd of the exist routines, implement the simulation in MATLAB.
Implement AHP to achieve best human satisfaction, optimize the searching algorithm with human preference factors and build a searching system interface demo in MATLAB
Feb, 2008
SUDOKU Game design (MATLAB, SAS)
The Mathematical Contest of Modeling (USA)
Implement math equations of Sudoku rules in MATLAB.
Define the concept of “known information amount” to setup unknown cells of a game and control the difficult level.
Using Principal component analysis (PCA) in SAS to analyze human satisfaction with games and feeling about difficulty
Create random game generate interface as a demo of the model.
Reading
“Options, Futures and Other Derivative Securities” John Hull
Honors and Awards
2008 Honorable Mentioned Award, Mathematical Contest of Modeling(MCM), USA, team leader
2007 Honorable Mentioned Award, Interdisciplinary Contest of Modeling(ICM), USA, team member
2007 Second Award, National Mathematical Contest of Modeling, China
2007 AUO Company Scholarship, Awarded to top 5%
2005 Excellent Student of the University
2004 Excellent Student of the University