While the second semester is in the starting blocks, I finally got some time to write something on the project I did in the first semester, for my Digital Signal Processing ‘Advanced topics’ course. The project was about vision-based pose tracking in the Wearable Augmented Reality (WEAR) program.

Augmented reality consists of adding a layer of information to what you’re seeing, for example keeping you up to date with online reviews of the restaurant you’re standing in front of. For a number of tasks that require many instructions, procedures or manuals, having this overlap can greatly increase efficiency, as proved in the ARMAR project of the University of Columbia and the US Marine Corps.

In the WEAR project, Belgian astronaut Frank De Winne has been testing a system on the ISS designed by Space Applications Services for the European Space Agency. For this system, the ‘pose’ – where the user is looking – is determined by comparing video data with a 3D model of the environment. This requires some advanced image processing wizardry, and this part was outsourced to the Electrical Engineering department at my university. I got the opportunity to play around with the software they made, and tested the system in a couple of environments it wasn’t really made for. You can see some results in the videos underneath. I would also like to thank David Tingdahl for helping me out with all of this.