Computer Vision for Artists and Designers: Pedagogic Tools and Techniques for Novice Programmers

Four Detection and Tracking Methods:
- Detecting Motion (frame differencing)
- Detecting Presence (background subtraction)
- Detection Through Brightness Thresholding (pixel brightness is compared to threshold value)
- Simple Object Tracking (locate & track brightest/darkest/coloured pixel)

There are three challenges with the current vision technology, and they are: lighting, quality, and distance. When using any type of camera, you need to have some type of light to produce any type of image. There are a few issues you can run into, such as participants wearing similar coloured clothing as the background. A few things a programmer could do, is having the participant stand in front of a light and then shine the light onto a screen in front of the camera. This way the participant’s silhouette is seen on the screen and it makes it easier for the camera to pick up the difference between light and dark. You can also create a controlled environment so that the camera does not pick up on any distractions.
The second challenge is quality. The quality of the camera is important; if it cannot pick up a lot of detail than you cannot do very precise measurements. Some ideas to help overcome this problem: a high quality camera, focus/zoom in on specific areas, use colour tracking, and infrared lights (to help reduce noise).
Cameras are limited in the distance that they can read. You need to limit where the participants can go, get a lens that has a large depth of view, and how much room they will need to perform. Physical interactions must be limited to a specific area or the camera will not read the movements.