Modern mobile phones are power houses of Visual Computing. For visual output, they provide a high-quality color display, and some models have dedicated hardware for 2D and 3D computer graphics. For visual input they have good digital cameras, both still and video. And they provide a fast CPU, some models even an FPU, as well as C++, Python, and Java programming environments. The Visual Computing and User Interfaces (VCUI) team studies how these visual capabilities can be harnessed for exciting applications.
One area we work on is mobile augmented reality. Using hints of location from a GPS or other signal sources (cell tower ID, WLAN station, Bluetooth beacon) and real-time image processing we want to build systems that recognize where the device and the user are, and what they are looking at. Based on recognized target the user can be given additional information in form of text, web links, overlayed graphics, etc. Applications range from personal navigation guide to tourist guide to any location-based services to outdoor and table top gaming.
We also work on mobile imaging. Camera phones are a good platform for computational photography. With the same device you can take several images, and then process them to create new images that you could take with a single shot using a normal digital camera. Instead of waiting until you can later process the images on a PC, we'd like to provide the instant gratification of seeing the results when you still are interested in the topic and can take additional input images if needed.
Our team develops and maintains NokiaCV, a computer vision library for S60 smart phones.
We are open to collaboration, our partners include Profs. Bernd Girod, Leo Guibas, and Marc Levoy (Stanford), and Gaetano Borriello and Steve Seitz (University of Washington), and many others.
