Advances in Computer Vision - Vol. 1

The series Advances in Computer Vision has the goal of presenting current approaches to basic problems that arise in the construction of a computer vision system, written by leading researchers and practitioners in the field. The first two volumes in the series comprise seven chapters, which together cover much of the scope of computer vision. Chapter 1 of Volume 1 is referred to as 1.1, and so forth.

In these volumes, computer vision means computer programs analyzing visual input (like a television image of a three-dimensional scene) and deriving from the image some description of the scene that is helpful to further reasoning or action concerning the scene ( Pentland, 1986). The technical aspects of creating images and transferring them to the computer's memory are not addressed here: We are concerned with the techniques of computerized image analysis. Computer graphics, or the computerized production of images, is the inverse of computer vision: Graphics starts with world descriptions and produces images. Since graphics is so closely related to vision, some graphics techniques are indeed interesting to vision researchers (see chapter 2.4).

Vision is our most powerful sense. If we could endow machines with something like the power of sight, their actions and decision-making could be much more effective and efficient. Automatic navigation systems, industrial inspection, biomedical applications, consumer products, interactive systems, and a host of other areas are potential beneficiaries of breakthroughs in computer vision. Of course in a real system vision is purposive, working toward some particular goal. That means that vision

-ix-