Student research opportunities
Build a stereo vision system and develop optimal stereo matching algorithms
Project Code: CECS_625
This project is available at the following levels:
Honours, Summer Scholar, Masters, PhD
Keywords:
Computer vision, image processing, 3d modelling
Supervisor:
Dr Hongdong LiOutline:
Our human eyes can perceive the 3D geometry of our world effortlessly. This is largely due to our two eyes--which form a binocular stereo system---that offer an almost perfect solution of 3D sensing.
In this project, you will asked to study and develop a desktop stereo vision system consisting of two synchronized fire-wire (800M) cameras. More importantly, you will be doing literature reading (into most recent, and top Computer Vision Conferences papers such as that in ICCV/CVPR/ECCV) and based on which to research a new, and optimized stereo-matching algorithm to compute a 3D depth-map which is not only fast, but also accurate.
Goals of this project
At the completion of this project, you will understand how the graphical 3D TV scenes are filmed, and grasp basic and fundamental image processing and computer vision knowledge and skills.
You may also be well-prepared for pursuing a PhD or Research Master degree in computer/robot vision or computer graphics.
Requirements/Prerequisites
Linear algebra
Matrix computation
Knowledge of Matlab and its image processing toolbox.
Basic C/C++ programming.
Student Gain
Working with top professors and researcher at ANU and/or NICTA.
Background Literature
Any computer vision textbook.
