Possible Project Ideas

Before you read any further, please be aware that this is just a list of things that I would like to do if I had the time. I'm not saying they are all suitable, and most importantly - I am not able to supervise any of these projects. You would need to find your own supervisor.
Ideas are added as I think of them or when I feel the urge to update this page. Feel free to send me an e-mail if you want to discuss these ideas further.

• IPv6 on Linux.
  We have a couple of labs of computers running Linux which are used in teaching the Robotics / Intelligent Active Vision papers. These computers are separated from the rest of the campus network by a router, and it would be interesting to install IPv6 on them. This would need to run in parallel with the current IPv4 stack as these computers are used throughout the year for teaching purposes. Initially this would be implemented locally, but if the results were promising then we could consider connecting to the 6bone.
  
• Dynamic Digital Watermarking on the Web
  Breach of copyright is very widespread on the internet, with no viable solution in sight. Digital watermarking allows images to be invisibly tagged to track their movements. If images displayed on a web site were tagged with a unique serial number each time they were sent, and this serial number was logged with the connection information and possibly some other details about the requester, when those images turned up on another site the user who downloaded them could be traced.
  Ideally this would be implemented as an Apache module. The watermarking technique would need to be very efficient, but not particularly robust.
  
• SNMP monitoring of our labs
  The CITR maintains three graduate computing labs with a large and diverse assortment of computers, some of which dual-boot their operating systems. I would like to use my web browser to graphically monitor the status of our computers, including information about what operating system is running, the load average, and who is logged on. Secondly, I would like to have reports generated which summarize this information over time.
  The University of Waikato has a similar system which could form the basis for this project
  
• Image Database
  Prof. Klette has been interested in the idea of an image database for some time now. Some work has already been done in this area by previous project students. My interest lies in possible implementations of the database back-end for storing and retrieving images and associated meta-data. Using an Object-Relational database such as PostGreSQL would allow data types to be defined for storing images, and could implement methods to manipulate and compare those images. Moving the majority of computation to the server would allow for a simple web-based client that would not require a Java applet to be run.
• Representing and Summarizing Qualitative Survey Data
  Existing relational databases can easily represent quantitative survey results such as multi-choice questions where a respondant is asked to rate their views on a scale, or answer Yes/No. Where these systems fail is in capturing comments that a respondant may be asked to make to justify their response, and often this is where the most valuable information lies. Extensions to an Object-Relational database could allow this data to be captured and analysed for common themes. One simple approach could be based on identifying keywords or common phrases but other, more sophisticated, techniques could be investigated. .