What we do
There are many excellent resources to help you decide whether or not to do graduate work. A PhD is a particularly testing endeavour, and it isn't for everyone. You should think very seriously about what you want to achieve before embarking on the graduate student journey.
Do not be fooled into thinking that image processing is an "easy" field. This is not an uncommon perception, but success in the field requires a demanding combination of mathematical, analytical and programming ability. It is also a fun and satisfying area to work in, so I would certainly not want to discourage you. However, if you are a weak student who just wants to get a graduate degree, then this is likely not the field for you.
You are obviously not expected to know everything at the start and the precise details of what you need to know will depend on the specific project. To give a rough idea though, students starting out in our research group will benefit from knowing the following:
- (Necessary) An ability to program competently with matlab and/or python. Experience with relevant software libraries is useful, but not at all necessary;
- (Necessary) A solid understanding of at least basic statistics;
- (Necessary) A solid understanding of linear algebra to the level of at least eigenanalysis. Knowledge of SVD would be useful;
- (Necessary) Good English skills. Writing a thesis is hard for everyone, and if you barely meet the minimums required for entry in the University's programmes then you will need to undertake significant study to improve your skills;
- (Necessary) Exposure to state space ideas (or phase space if you have a physics background);
- (Preferred) Basic digital signal processing knowledge, such as comfort with the Fourier domain and familiarity with topics such as sampling and digital filtering;
- (Preferred) Some exposure to estimation methods such as the Kalman filter;
- (Preferred) Some knowledge of machine learning;
- (Preferred) Some experience with image processing.
I should emphasise that previous experience with image processing is not at all necessary. In my experience it is much easier to pick up the specifics of working with images than it is to learn the underlying notions of how to work with general signals. there are of course a wide range of interesting things to learn, butiIf you have a good set of tools then the application to vision systems should be relatively painless.
The requisite background skills are most likely to be possessed by students with an undergraduate background in electronic engineering, experimental physics, statistics, or computer science with machine learning
. A generic computer science background without significant mathematical or statistical exposure is unlikely to provide good preparation to join our group.
Masters students could choose to work more at the hardware level, where knowledge of electronic and mechanical system design would be advantageous, along with an ability to program in python and matlab. Mathematical or statistical background may not be as important for such students, depending on project choice. Again, feel free to get in touch if your interests are in this area.
If you are interested in joining the group, then please do the following:
- Visit the website of the Faculty of Graduate Research for information on applying for graduate research and for details of the procedures at VUW.
- If you are a local student, arrange to come and have a talk about options. Otherwise, send me an email outlining your interests, along with transcripts for tertiary studies that you have completed and a complete CV. If you don't provide this information, I will just need to ask for it.
- Let me know what it is that actually interests you in doing research work. What will get you out of bed in the morning? Doing fundamental mathematical work, or implementing existing algorithms? Do you want to design mechanical devices, or electronic circuits? Are you interested in studying the intricacies of some piece of code, or do you want to build a complete system that solves a real worl problem? Do you want to build things from scratch, or assemble something from pre-existing components? Is there some skill that you would really like to develop during graduate studies?
The more information you can give about what you actually want to do, the better I can do at matching you with a project that can satisfy your interests.