MIT EECS - Foxconn Undergraduate Research and Innovation Scholar
Track Visually Moving Objects Using RGBD Data
Tracking is an important topic within the field of computer vision and robotics. It includes some basic tasks of detecting moving objects and extracting information about their movement, such as individual’s position and velocity. The next higher level aim is to successfully track targets over time and analyze their behaviors from the tracks obtained. The goal of this project is to develop a new method to track moving objects from video images sequences. The approach that we would like to use is non-parametric Bayesian inference. The reason is that we don’t have much information about the priors and would like to do inference continuously online. And non-parametric models could automatically infer adequate model complexity from the data
I have worked at MIT Media Lab on constructing 3D model of an object by aligning point clouds obtained from a depth camera. As an intern, I worked on a data visualization project at Oracle. I also have experience building an Android application which serves as a user interface for people in a treatment center.