Computer Vision
The Camera on the OLPC XO has vast potential beyond that which is exposed in Record. By using computer vision, it can be used to provide a more interactive and immersive user experience.
Pygame
Pygame will soon have support for cameras and some computer vision functions. This is likely the easiest way to integrate vision with Activities.
While still under heavy development, source for the branch of pygame with camera support and vision functions can be checked out from: [1]
Of particular note are the usage examples here and here.
Use Cases
Accessibility
Gaming
Currently, the only game taking advantage of vision processing is Colors!, which allows you to use the XO's bright green charger as a virtual paint brush. It accomplishes this by converting the colorspace of a frame from a gstreamer buffer to HSV and finding the centroid of the specific hue of the charger. This is a fairly complicated process that will soon be made much simpler upon completion of camera support for Pygame, a GSoC 2008 project.
Physics
See Physics#Physics_and_vision_processing
Art
See Living Pointillism.