Work-Study Projects Available at iLab

Warning: For your contribution to be useful to us, it is important that you truly be in charge of the project you chose, so that we spend less time with you than we would by doing the projects ourselves. These projects hence are to be considered as work projects, not student research projects (since we usually spend a lot more time with students pursuing a PhD in our lab, as part of our education mission towards these students). So if you are interested, it will be essential that you are ready to figure out everything that needs to be figured out, mostly on your own after we give you a brief introduction to the projects and all the pieces of information we have so far. That means using google, calling people, doing research on how others have solved similar problems, etc.

We will only hire students who can work under a work-study contract where the University participates to the expense. Please make sure that this applies to you!

iLab Publication Web Site

On our publication web site only about half of our publications are online as PDFs. We need to find a copy of, and scan-in, the others. This will require web research for the final published version of each paper (if a PDF is available online on the publisher's web site), and the use of a scanner to digitize and convert to PDF the papers which do not exist in electronic form at all. You will then edit our master publication database to put in all the details about each publication you have scanned (mostly, only the abstract and link to a PDF file should be missing from the database).

iLab Publication Server Engine

Our publication web pages are currently generated by a Perl script that takes a BibTeX file as input. We wish to use a database instead. We have preliminary Perl code to query a PostgreSQL database and generate similar pages. But it is not finished. This is all Perl/Linux development. See here. There will be no Windows machine or software involved in this project, this is 100% Linux.

Image Databases

We have several image databases but they need to be organized and cleaned up. This is all Linux too. Writing scripts to move/rename/convert large numbers of image files. Our goal is to have all images in 24-bit PPM format and possibly with some standardized file name convention. See here for our current collection of databases that have already been cleaned up and packaged.

Images

We need more images, and we have a digital camera. The images we need are to put on the web, showing our various research projects, some demos, etc. So your role there would be like a public relation person: learn about the projects in the lab, ask the students to give you demos, collect screen grabs and information from them. Generate cool web pages. Similar to, for example, here which summarizes one of our papers.

Video Games

We think our vision algorithms described here can play them. Task: find a way to digitize video from a video game with high quality. For example, play on one PC, then use a VGA-to-NTSC converter, then digitize the NTSC on another PC (pretty overkill but what we want is to grab the game at 30 frames/s). Then run our visual attention algorithm. Generate a report similar to that generated by an Art School student working with us here.

Refine Beobot web site

By adding various cool features such as photos, screenshots, movies, etc. See here. The server is Apache under Linux.

Microcontroller Programming

We wish to consolidate the various peripherals (control of servos, lcd screen, small keypad, various analog & digital sensors) of our Beobots into a unique PIC-based design. Requires experience with assembly code programming of microcontrollers such as a PIC 16C74B from MicroChip.