If you're looking for a project to work on for your bachelor's or master's thesis, here are some specific projects I would love to supervise. Please get in touch if you would like to hear more.

I'm also interested generally in interaction design, pervasive computing, as well as mobile and social computing. Feel free to suggest something as well.

• Live post-it: A camera/projection system that can identify the position and rotation of a regular post-it notes on a whiteboard. Coordinates are transformed and mapped to screen coordinates in order to overlay a post-it with live data from a static projector. Algorithm should scale to identify multiple, non-overlaid post-its.
• Feasibility of vision-based physical token authentication between a webapp and mobile phone. The mobile phone generates and displays a series of patterns. User visits a website on her computer and logs in by holding the phone’s display up to the computer’s camera. Image processing should be executed in client-side Javascript running in the browser.
• Feasibility of vision-based tracking by a mobile phone of a screen image. Based on a screen-camera feedback loop, can a system estimate the position, rotation and distance of a mobile phone (with camera faced toward screen) in relation to the screen. Ideally using marker-less tracking. Phone can thus be used as an extension of screen surface or controller of screen.
• Wearing Data: Explore the idea of wear on digital artifacts. A mobile app is loaded with a file (picture/music/video). The app degrades the file over time based on the user’s activity, using for example sensors embedded in the device and system status (battery, radio coverage etc). User has no direct control over how sensor input impacts capture, yet data processing happens in a consistent way so that the user might learn how to influence the process to produce aesthetically-pleasing results. The file should be corrupted (or “glitched”), yet still openable for viewing or playback.
• Digital Lab: Create a mobile app for photo/video/sound capture that uses the current of the phone as an input into the capture process. For example, based on on-board sensors, battery level, radios etc. User has no direct control over how sensor input impacts capture, yet data processing happens in a consistent way so that the user might learn how to influence the process to produce aesthetically-pleasing results. State of phone immediately following capture could also be combined in the “developing” of the file.
• Skin Cancer Prevention: Create a mobile app/website that uses the phone’s camera to track whether your skin is burning (for example when you’re at the beach for a day) and to advise you when to reapply sunscreen or to get out of the sun altogether. Image processing is combined with geolocation, ambient light sensor, weather & UV conditions to give safe suggestions.
• Web-based distributed data store: Client side storage is now available in most web browsers. Exploit this to provide distributed storage and retrieval of data without requiring installation of additional software at storage endpoints. Files can be reliably reconstructed from distributed blocks should the original source become unavailable by the coordinated action of users operating the client machines. Ideally clients should should have plausible deniability and store only opaque, encrypted blocks of data.