The Microsoft Kinect uses IR to capture a depth image of its subject, a person, in our case, and its camera to get the colours. This data is then sent to the computer where the Processing app visualises it as a 3D point cloud of voxels (three-dimensional pixels). This point cloud is what the audience can see on the screen.
As for the interactive part, the app allows an audience member to "take a selfie" in this three-dimensional space. After entering an email address, a virtual camera rotates around the point cloud while recording an animated GIF, thus giving a sense of space and relative movement. The participator will then receive the animated GIF by email.