Research Assistant @ Computer Vision Lab (SBU)
- Topic: Human action recognition in videos
- proposed a method to recognize and prune non-action shots for action recognition
- proposed a method to separate human actions from dominant context factors
Undergraduate Dissertation @ Information Processing Center (USTC)
- Topic: Image Super Resolution Reconstruction.
- Modeled the imaging process with a Bayesian network that encompasses optical flow, blur kernel, down-sampling and noise;
- Derived estimating formulations under MAP principle.
- Solved optimization using IRLS (Iterative Reweighted Least squares) and CG (Conjugate Gradient) method.
Research Assistant @ Information Processing Center (USTC)
- Topic: Image Processing and Analysis
- Practiced on image processing programming using Matlab/C++.
- Learned about matrix analysis, statistical learning and convex optimization.
- Worked on a prototype of Content Based Image Retrieval System.
National University Student Innovation Program
- Topic: miRNA Target Prediction using Machine Learning Algorithm.
- Constructed comprehensive mRNA-miRNA dataset by parsing mirTarbase webpage and processing Microarray data.
- Utilized mRMR method to select optimal feature set from gene functional information and sequence information.
- SVM performance achieves higher accuracy, meanwhile selected GO features appropriately represent gene functional information.
USTC-MSRA Smart Campus Program
- Topic: USTC Social Network Construction.
- Built a social network site based on UCHome, redesigned certain interface and functionality.
- Upgraded the library news site from DedeCMS to WordPress.
Mathematical Contest in Modeling (MCM) 2013
- Topic: Water Strategy Analysis of China
- Considered 4 strategies: Water Movement, Desalination, Conservation and Market-Oriented Approach.
- Introduced a mathematical model with optimizing goal and constraints.
- Utilized mRMR, PCA and Metropolis Algorithm to optimize for water equilibrium.
Robot Game Contest 2011
- Topic: Human-assisting Robots
- Built a robot that assists blind people to walk at home and on street.
- Our robot relies on assorted sensors (ultrasonic/infrared/color sensors).
- Our robot can be controlled with voice commands using speech recognition chip(Lingyang-SPCE061A).