Computerized visual inspection systems are increasingly being considered and implemented as an alternative to manual inspection and manual gauging in manufacturing systems. Although computer vision systems and image processing have received much attention in the past, statistically powerful, yet computationally efficiency methods still remain to be developed to process the visual information to make computerized inspection systems cost effective. This study investigates the case of inspecting round production parts for shape and size. The statistical properties of the out-of-roundness measurement procedure are investigated. Real-time implementation considerations are discussed.