Interactive Visual Prototyping of Computer Vision Applications

Cameras are a useful source of input for many interactive applications, but computer vision programming requires specialized knowledge that is out of reach for many HCI practitioners. My thesis is that by using visual examples to train computer vision classifiers, and combining trained classifiers into graphs, people with minimal programming experience, and no specialized knowledge of computer vision, can quickly and easily build a wide variety of useful camera-based applications. I will demonstrate this thesis through the design and evaluation of a prototyping tool for camera-based interaction, which I will build, distribute, evaluate, and incrementally refine over the course of the coming year.