Fast techniques in object, edge, and face detection

In the quest to provide computers with the understanding of the world around them, much work has been done in the fields of computer vision, robotics, and image processing to provide the computer with tools to understand our world in simple terms. With these tools, researchers have attempted to coax the computer to perform complex tasks which we have taken for granted. A human face, for example, is an object even the youngest child has no problem in locating, but locating a face is not a natural task for a computer.; We present in this thesis algorithms developed for some of these tasks with special interest in keeping the algorithms fast for a computer to perform. A typical computer is largely a serial computer executing instructions one at a time in sequence. By keeping the algorithms simple and fast, they are more likely to be used as building blocks in a more complex task. We introduce algorithms that rotate an image quickly, reduce the environmental lighting effect on an image, and find certain kinds of edges in a noisy image.; We also present solutions to the complex task of face detection. We will describe a solution when a video camera is in a controlled environment and when one is provided with only a single image. What are presented are not, however, intended as full solutions. They are still meant as building blocks for a larger system; one can view them as pre-processors. They are primarily meant to ease the computational burden of whatever algorithm follows it.