A closer projection silhouette

It's weird how difficult a title can be when you are trying to do your thesis. The spine of the print needs this many letters, the title online needs this much clarity, and somewhere there is an unwritten rule that you can't just use the word "stuff".

My Masters of Science is from the University of Calgary, in a program called Computational Media design. It is intended to be a bridge between Art, Design and Science. As an interdisciplinary degree this includes several faculties, including Art, Computing Science, Environmental Design Studies and Music. It's fairly broad. My Masters was in EVDS and Computing Science, mostly, but I did manage to get a few courses in 3d modelling from both EVDS and the Faculty of Fine Arts. I learned several techniques which helped me with my thesis project.

For my thesis I was concerned with trying to get a viewing audience closer to a projection screen. My goal was to be able to create a realistic silhouette of a person while they were within 6 feet of the surface of a projection screen. While doing so I managed to be able to allow touch and hover interactions with the surface, and to be able to determine individual people within this kind of space. I did this using multiple Microsoft Kinects as depth cameras.

You can see more of what is involved within this video, prepared as part of the defense presentation for my thesis. Long story short, I reconsidered the Kinect as a lighting source, instead of a camera, and used openGL for placement and to resolve where points were within the bounds of the projected screen. The touches are fairly rudimentary, as is the method for determining which points belong to individuals, but the prototype worked. I was able to generate these images while standing within 5 feet of the screen, and this system will allow for the completion of some artwork intended to enhance a personal experience for those who approach the screen.

I also explored several methods for displaying information regarding the individual points. In this image you can see I have used both color to display depth, as well as FatFonts, a method for displaying large numbers through changing line thickness and embedded values.

This method also allows the interaction of multiple people within the space, and alternatively could be used to gather information all the way to the floor. displaying it on the monitor available resulted in a somewhat squished image, but you can see the feet of those within the space.

The biggest personal development came from devising a way to place the Kinects in 3d space in relation to the screen. I had created a similar system with 6 Kiencts while at ACAD on a much larger screen, but that system was not easy to set up and, after I left, became impossible for others to maintain. This portion of my project is still under development, but eventually will be made available so that others may explore closer interactions for viewers within their own artworks and media.