Electronics and semiconductor manufacturing are the two top markets for machine vision products, accounting for an estimated $620.4 million in North American machine vision revenues in 1998, according to the Automated Imaging Association. Because they account for such a large share of the $1.5 billion North American market, electronics and semiconductor manufacturing have strongly influenced the development of advanced machine vision technologies that can also benefit many other industries. The reasons for their strong and early reliance on machine vision relate to the difficulty and importance of the quality assurance role. Products are small and thus difficult and/or time-consuming for human inspectors to assess. On the other hand, profit margins are so tight that it's important to assess partially complete products between each manufacturing step to avoid adding value to something that is already deficient. In addition, it's important to catch and fix manufacturing problems early to maximize yields and plant efficiency. The overarching goal of machine vision in electronics manufacturing, then, is a standard business goal: Make it better (smaller, more reliable, more functional), faster and cheaper. The challenge for machine vision is to produce technologies that work not only with whatever is small, fast and cheap today but also the next generation of products, which will be even smaller, faster and cheaper tomorrow. Over the last five years, machine vision has met the challenge by reducing costs, simplifying the human/machine interface and improving accuracy and resolution. In the next five years, it will build on those improvements and add several new technologies 3-D vision, color capabilities and "smart" systems -- to solve problems that customers are only beginning to notice.