Is Computer Vision Still AI?

Bayesian); and nonsymbolic image-toimage transformations. Most successful practical vision systems are wellengineered combinations of specialpurpose sensors, hardware, and algorithms that are tailored to solve specific visual problems (and are generally more successful in proportion to the narrowness of the task). Some texts (for example, Batchelor, Hill, and Hodgson [1985]) don’t even mention AI or its methods. Example practical applications include assembly-line robot welding guidance, inspection of computer keyboards and integrated circuits, counting of tea bags, tomato grading, and optical tracking. We can contrast computer vision’s rather specific research directions with the outward appearance of AI, ■ Recent general AI conferences show a decline in both the number and the quality of vision papers, but there is tremendous growth in, and specialization of, computer vision conferences. Hence, one might conclude that computer vision is parting or has parted company with AI. This article proposes that the divorce of computer vision and AI suggested here is actually an open marriage: Although computer vision is developing through its own research agenda, there are many shared areas of interest, and many of the key goals, assumptions, and characteristics of computer vision are also clearly found in AI.